Solanum identifier using Neural Networks

Contains the code to balance classes and create the training/val/test sets

Environment preparations

Dependencies

# For Python 3.7
#!pip install keras tensorflow matplotlib numpy pandas imblearn split-folders scikit-learn tensorflow-addons scikeras[tensorflow]

# For Python 3.10
#!pip install "keras==2.10.0" "tensorflow<2.11" matplotlib numpy pandas imblearn split-folders scikit-learn "tensorflow-addons<2.11"
#!pip install --no-deps keras-tuner

Env vars

# Set this variable to the root path of where the files are located
%env DATA_LOCATION=P:\CODE\ITESM\tesis-dataset-downloader\solanum_output
%env CLEAN_DATA_FOLDER=z_clean_resized
%env CLEAN_SPLIT_DATA_FOLDER=z_clean_resized_split

# Allow an async memory allocation to avoid OOM errors when using multiple models
#%env TF_GPU_ALLOCATOR=cuda_malloc_async
%env TF_FORCE_GPU_ALLOW_GROWTH=true

!echo %DATA_LOCATION%

#!dir %DATA_LOCATION%

# Constants
import os

RANDOM_SEED = 1988
DATA_ROOT_LOCATION = os.environ["DATA_LOCATION"]
CLEAN_DATA_FOLDER = os.environ["CLEAN_DATA_FOLDER"]
CLEAN_SPLIT_DATA_FOLDER = os.environ["CLEAN_SPLIT_DATA_FOLDER"]

env: DATA_LOCATION=P:\CODE\ITESM\tesis-dataset-downloader\solanum_output
env: CLEAN_DATA_FOLDER=z_clean_resized
env: CLEAN_SPLIT_DATA_FOLDER=z_clean_resized_split
env: TF_FORCE_GPU_ALLOW_GROWTH=true
P:\CODE\ITESM\tesis-dataset-downloader\solanum_output

Imports

%matplotlib inline
import tensorflow as tf
from tensorflow.keras.utils import img_to_array

from tensorflow.keras.preprocessing import image
from tensorflow.keras.preprocessing.image import ImageDataGenerator

import tensorflow_addons as tfa

import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import ImageGrid

import numpy as np
from numpy import array

import pandas as pd

from imblearn.under_sampling import RandomUnderSampler
from imblearn.over_sampling import RandomOverSampler

import splitfolders

from sklearn.metrics import classification_report, confusion_matrix, ConfusionMatrixDisplay
from sklearn.model_selection import KFold, ShuffleSplit

from PIL import Image

import os
import time
from datetime import datetime

import random

# Enable 3rd Party Jupyter Widgets in Google Collab
#from google.colab import output
#output.enable_custom_widget_manager()

def get_training_device_name():
    name = tf.test.gpu_device_name()

    if "GPU" not in name:
        print("No GPU was found!, training will be done in the CPU which will be slower")
        name = '/cpu:0'
    else:
        print('Found GPU at: {}'.format(name))
    
    return name

TRAINING_DEVICE_NAME = get_training_device_name()

random.seed(RANDOM_SEED)
tf.keras.utils.set_random_seed(
    RANDOM_SEED
)

Found GPU at: /device:GPU:0

Tools

def print_marquee(msg: str):
    """
    Prints a centered message with a marquee of * 
    """
    marquee_width = len(msg) + 4
    print("\n")
    print("*" * marquee_width)
    print(f"* {msg} *")
    print("*" * marquee_width)

def count_rows_by_column(df, column):
    return df.groupby([column]).size().reset_index(name='count').sort_values("count", ascending=False)

def model_2_pkl(model, filename: str):
    """Exports a model to PKL format"""
    pickle.dump(model, open(filename, 'wb'))

def pkl_2_model(filename: str):
    """Loads a model from a PKL file"""
    return pickle.load(open(filename, 'rb'))

def get_class_names(generator) -> list:
    return [k for k, _ in sorted(generator.class_indices.items(), key = lambda item: item[1])]

def generate_id() -> str:
    # Generates an ID in the shape of YYYYMMDDhhmmss to be used in models and files
    return datetime.today().strftime('%Y%m%d%H%M%S')

def show_images(df: pd.DataFrame,
                labels_col_name="section",
                rows=2,
                cols=2,
                samples_per_label=1,
                figsize=(15,15)):
    
    sections = np.sort(df[labels_col_name].unique())

    for section in sections:
        parent_fig = plt.figure(section, figsize=figsize)
        plt.title(section)
        plt.axis("off")

        samples = df[df[labels_col_name]==section].sample(samples_per_label,
                                                          random_state=RANDOM_SEED)
        samples = [Image.open(path) for path in samples["full_path"]]
    
        grid = ImageGrid(parent_fig,
                         rect=111, # This will set the axes position as (left, bottom, width, height) tuple or as a three-digit subplot position code, it’s 111 in our case
                         nrows_ncols=(rows, cols),  # creates rows,cols grid of axes
                         axes_pad=0.1,  # Padding between pictures
                        )

        for ax, im in zip(grid, samples):            
            ax.imshow(im)
            ax.axis("off")

        plt.show()

Data preparation

Data Loading

images_df = pd.read_csv(os.path.join(DATA_ROOT_LOCATION, CLEAN_DATA_FOLDER, "images_dedup_512x512_100picspersection.csv"))

# Update all paths to conform to the local structure (Only if running in a UN*X environment)

# 'P:/CODE/ITESM/tesis-dataset-downloader/solanum_output/z_clean_resized/acanthophora/acanthophora_acerifolium_1928496814_gbif_2700.jpg'
# path_to_replace = "P:/CODE/ITESM/tesis-dataset-downloader/solanum_output/z_clean_resized"
# images_df["full_path"] = images_df["full_path"].str.replace(path_to_replace, data_root_location)

display(images_df.describe(include="all"))

	section	species	filesize_mb	image_type	source	full_path
count	8515	8515	8515.000000	8515	8515	8515
unique	12	112	NaN	1	2	8515
top	petota	lanceolatum	NaN	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...
freq	1733	533	NaN	8515	6103	1
mean	NaN	NaN	0.029506	NaN	NaN	NaN
std	NaN	NaN	0.005773	NaN	NaN	NaN
min	NaN	NaN	0.012600	NaN	NaN	NaN
25%	NaN	NaN	0.025500	NaN	NaN	NaN
50%	NaN	NaN	0.029200	NaN	NaN	NaN
75%	NaN	NaN	0.033200	NaN	NaN	NaN
max	NaN	NaN	0.062900	NaN	NaN	NaN

Data at a glance

print_marquee("Classes")
count_per_section = count_rows_by_column(images_df, "section")
display(count_per_section)

display(count_per_section.describe())

print_marquee("Info")
display(images_df.info())

***********
* Classes *
***********

	section	count
9	petota	1733
5	holophylla	1449
7	melongena	1442
11	torva	1332
2	brevantherum	811
10	solanum	577
3	dulcamara	389
4	herposolanum	224
8	micracantha	180
6	lasiocarpa	178
0	acanthophora	100
1	anarrhichomenum	100

	count
count	12.000000
mean	709.583333
std	617.223027
min	100.000000
25%	179.500000
50%	483.000000
75%	1359.500000
max	1733.000000

********
* Info *
********
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 8515 entries, 0 to 8514
Data columns (total 6 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   section      8515 non-null   object 
 1   species      8515 non-null   object 
 2   filesize_mb  8515 non-null   float64
 3   image_type   8515 non-null   object 
 4   source       8515 non-null   object 
 5   full_path    8515 non-null   object 
dtypes: float64(1), object(5)
memory usage: 399.3+ KB

None

Splitting the test set

Given that we will sample/undersample our dataset, we need to remove 10% of the classes for testing so that we don't contaminate our results and have the model train on test samples, the rest that we get will be balanced.

# Iterate over each of the classes and use train test split to sample out 20% of each class
# put this 90% and 10% into new DataFrames

TEST_FRACTION = 0.1

def split_balanced_dataset(original_images_df, test_fraction=TEST_FRACTION, is_verbose=True):

    training_set = pd.DataFrame()
    test_set = pd.DataFrame()

    for section in original_images_df["section"].unique():
        images_for_section = original_images_df[original_images_df["section"] == section]

        test_set_section = images_for_section.sample(
            frac=test_fraction,
            random_state=RANDOM_SEED,
            replace=False
        )

        # Remove the test indices from the train set
        training_set_section = images_for_section.drop(test_set_section.index)

        if is_verbose:
            print_marquee(section)
            print(f"Before sampling: {images_for_section.shape}")
            print(f"Training set for section: {training_set_section.shape}")
            print(f"Test set for section: {test_set_section.shape}")

        training_set = pd.concat([training_set, training_set_section])
        test_set = pd.concat([test_set, test_set_section])

    # Shuffle the training set to avoid issues with CV later
    training_set = training_set.sample(frac=1, 
                                    random_state=RANDOM_SEED).reset_index(drop=True)
    
    return training_set, test_set

training_set, test_set = split_balanced_dataset(images_df)

print(f">>>Train DF: {training_set.shape}\n   Test DF: {test_set.shape}")

****************
* acanthophora *
****************
Before sampling: (100, 6)
Training set for section: (90, 6)
Test set for section: (10, 6)


*******************
* anarrhichomenum *
*******************
Before sampling: (100, 6)
Training set for section: (90, 6)
Test set for section: (10, 6)


****************
* brevantherum *
****************
Before sampling: (811, 6)
Training set for section: (730, 6)
Test set for section: (81, 6)


*************
* dulcamara *
*************
Before sampling: (389, 6)
Training set for section: (350, 6)
Test set for section: (39, 6)


****************
* herposolanum *
****************
Before sampling: (224, 6)
Training set for section: (202, 6)
Test set for section: (22, 6)


**************
* holophylla *
**************
Before sampling: (1449, 6)
Training set for section: (1304, 6)
Test set for section: (145, 6)


**************
* lasiocarpa *
**************
Before sampling: (178, 6)
Training set for section: (160, 6)
Test set for section: (18, 6)


*************
* melongena *
*************
Before sampling: (1442, 6)
Training set for section: (1298, 6)
Test set for section: (144, 6)


***************
* micracantha *
***************
Before sampling: (180, 6)
Training set for section: (162, 6)
Test set for section: (18, 6)


**********
* petota *
**********
Before sampling: (1733, 6)
Training set for section: (1560, 6)
Test set for section: (173, 6)


***********
* solanum *
***********
Before sampling: (577, 6)
Training set for section: (519, 6)
Test set for section: (58, 6)


*********
* torva *
*********
Before sampling: (1332, 6)
Training set for section: (1199, 6)
Test set for section: (133, 6)
>>>Train DF: (7664, 6)
   Test DF: (851, 6)

print_marquee("Training")
training_count_per_section = count_rows_by_column(training_set, "section")
display(training_count_per_section)

print_marquee("Info")
display(training_count_per_section.describe())
display(training_set.info())


print_marquee("Test")
test_count_per_section = count_rows_by_column(test_set, "section")
display(test_count_per_section)

print_marquee("Info")
display(test_set.info())

************
* Training *
************

	section	count
9	petota	1560
5	holophylla	1304
7	melongena	1298
11	torva	1199
2	brevantherum	730
10	solanum	519
3	dulcamara	350
4	herposolanum	202
8	micracantha	162
6	lasiocarpa	160
0	acanthophora	90
1	anarrhichomenum	90

********
* Info *
********

	count
count	12.000000
mean	638.666667
std	555.574123
min	90.000000
25%	161.500000
50%	434.500000
75%	1223.750000
max	1560.000000

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 7664 entries, 0 to 7663
Data columns (total 6 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   section      7664 non-null   object 
 1   species      7664 non-null   object 
 2   filesize_mb  7664 non-null   float64
 3   image_type   7664 non-null   object 
 4   source       7664 non-null   object 
 5   full_path    7664 non-null   object 
dtypes: float64(1), object(5)
memory usage: 359.4+ KB

None

********
* Test *
********

	section	count
9	petota	173
5	holophylla	145
7	melongena	144
11	torva	133
2	brevantherum	81
10	solanum	58
3	dulcamara	39
4	herposolanum	22
6	lasiocarpa	18
8	micracantha	18
0	acanthophora	10
1	anarrhichomenum	10

********
* Info *
********
<class 'pandas.core.frame.DataFrame'>
Int64Index: 851 entries, 77 to 8054
Data columns (total 6 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   section      851 non-null    object 
 1   species      851 non-null    object 
 2   filesize_mb  851 non-null    float64
 3   image_type   851 non-null    object 
 4   source       851 non-null    object 
 5   full_path    851 non-null    object 
dtypes: float64(1), object(5)
memory usage: 46.5+ KB

None

Balancing of classes

We can see that there are classes that are overrepresented, so we need to do some undersampling and oversampling in order for the model to better learn and classify.

The average count for the samples in the training set is 638.88 while the median is 434 (listed as the 2nd quartile or 50%), so the best strategy that will be employed is to randomly undersample anything above 500 and oversample anything below that level.

# As an experiment, we will train the model with just the pictures that have
#  at least 1000 samples to debug why we are not seeing the neural network learn
#  and ignore the sections that have less than that

#SAMPLE_SIZE = 1000
#sections_with_1000 =  list(training_count_per_section[training_count_per_section["count"] >= SAMPLE_SIZE]["section"].values)

#balanced_training_data = training_set[training_set["section"].isin(sections_with_1000)].reset_index()

#display(count_rows_by_column(balanced_training_data, "section"))

# https://imbalanced-learn.org/stable/

SAMPLE_SIZE = 1500

sections_to_oversample = list(training_count_per_section[training_count_per_section["count"] < SAMPLE_SIZE]["section"].values)
sections_to_undersample = list(training_count_per_section[training_count_per_section["count"] > SAMPLE_SIZE]["section"].values)

undersampler = RandomUnderSampler(sampling_strategy={k: SAMPLE_SIZE for k in sections_to_undersample},
                                  random_state=RANDOM_SEED)

oversampler = RandomOverSampler(sampling_strategy={k: SAMPLE_SIZE for k in sections_to_oversample},
                                  random_state=RANDOM_SEED)


def imbsample(imbsampler, df):
    """Applies a sampling strategy to a Pandas DataFrame using imblearn
    
    The imblearn only accepts the features and labels separately, so let's join
    the columns after the operations to get a single DataFrame
    
    :returns the df with sampled data 
    """
    X, y = imbsampler.fit_resample(df.drop("section", axis=1), df["section"])
    return pd.concat([X, y], axis=1)

balanced_training_data = imbsample(oversampler, training_set)
balanced_training_data = imbsample(undersampler, balanced_training_data)

# Shuffle the data
balanced_training_data = balanced_training_data.sample(frac=1,
                                                       random_state=RANDOM_SEED).reset_index(drop=True)

display(balanced_training_data.head(10))
display(count_rows_by_column(balanced_training_data, "section"))

	species	filesize_mb	image_type	source	full_path	section
0	hirtum	0.0315	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	lasiocarpa
1	erianthum	0.0270	jpg	idigbio	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	brevantherum
2	lanceolatum	0.0311	jpg	idigbio	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	torva
3	pubigerum	0.0234	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	holophylla
4	nudum	0.0367	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	holophylla
5	myriacanthum	0.0445	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	acanthophora
6	pubigerum	0.0309	jpg	idigbio	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	holophylla
7	refractum	0.0360	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	herposolanum
8	rostratum	0.0313	jpg	idigbio	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	melongena
9	oxycarpum	0.0430	jpg	gbif	P:/CODE/ITESM/tesis-dataset-downloader/solanum...	petota

	section	count
0	acanthophora	1500
1	anarrhichomenum	1500
2	brevantherum	1500
3	dulcamara	1500
4	herposolanum	1500
5	holophylla	1500
6	lasiocarpa	1500
7	melongena	1500
8	micracantha	1500
9	petota	1500
10	solanum	1500
11	torva	1500

Sample of images

show_images(balanced_training_data,
            rows=3,
            cols=3,
            figsize=(7, 7),
            samples_per_label=9)

Output hidden; open in https://colab.research.google.com to view.

Data augmentation

TARGET_SIZE = (224, 224)

# BATCH_SIZE = 32
BATCH_SIZE = 16
COLOR_MODE = "rgb"
X_COL = "full_path"
Y_COL = "section"

image_data_gen_params = {
    "target_size": TARGET_SIZE,
    "color_mode": COLOR_MODE,
    "class_mode": "categorical",
    "seed": RANDOM_SEED,
    "save_prefix": 'augmented_',
    "save_format": 'png',
}

def get_train_generator(preprocessing_func, training_df, batch_size=BATCH_SIZE):
    train_datagen = ImageDataGenerator(
                rescale=1.0/255, # Normalize the data to be 0-1
                rotation_range=40,
                width_shift_range=0.2,
                height_shift_range=0.2,
                shear_range=0.2,
                zoom_range=0.2,
                horizontal_flip=True,
                fill_mode='nearest',
                preprocessing_function=preprocessing_func)
    
    # return train_datagen.flow_from_directory(
    #   os.path.join(DATA_ROOT_LOCATION, CLEAN_SPLIT_DATA_FOLDER, "train"),
    #   # save_to_dir=os.path.join(DATA_ROOT_LOCATION, CLEAN_SPLIT_DATA_FOLDER, "train"),
    #   **image_data_gen_params)
    
    return train_datagen.flow_from_dataframe(
        training_df,
        x_col=X_COL,
        y_col=Y_COL,
        batch_size=batch_size,
        # save_to_dir=os.path.join(DATA_ROOT_LOCATION, CLEAN_SPLIT_DATA_FOLDER, "train"),
        shuffle=True,
        **image_data_gen_params)

def get_val_generator(df, preprocessing_func, batch_size=BATCH_SIZE):

    test_datagen = ImageDataGenerator(
        rescale=1.0/255,
        preprocessing_function=preprocessing_func,
    )

    return test_datagen.flow_from_dataframe(
            df,
            # save_to_dir=os.path.join(DATA_ROOT_LOCATION, CLEAN_SPLIT_DATA_FOLDER, "val"),
            x_col=X_COL,
            y_col=Y_COL,
            shuffle=False,
            batch_size=batch_size,
            **image_data_gen_params)

Model candidate training

We will perform Cross-Validation = 4 to find the best architecture and with it, we'll train again on all of the training set and proceed to validate it on the test set.

CV_FOLDS = 4
NUM_CLASSES = len(balanced_training_data["section"].unique())
VAL_SIZE = 0.20

TRAINING_RUN_ID = generate_id()

def train_model(model_to_fit: tf.keras.models.Model,
                fit_params: dict,
                preproc_func):

    t = time.process_time()
    print_marquee(f"Started at {t}...")

    batch_size = fit_params.get("batch_size", BATCH_SIZE)
    
    # Transform from DataFrames to File System Generator
    # Make a copy of fit_params so as not to edit the original dictionary
    fit_params = fit_params.copy()
    fit_params["x"] = get_train_generator(preproc_func,
                                          fit_params["x"],
                                          batch_size=batch_size)
    fit_params["validation_data"] = get_val_generator(fit_params["validation_data"],
                                                      preproc_func,
                                                      batch_size)
    
    trained_model_history = model_to_fit.fit(**fit_params)
    
    elapsed_time = time.process_time() - t
    
    print(f"\n\n ********* Training time: {elapsed_time} s.")
    return trained_model_history

def evaluate_model(trained_model: tf.keras.models.Model,
                   trained_model_history,
                   fit_params: dict,
                   preproc_func,
                   print_loss_graphs: bool=True,
                   print_model_info: bool=True) -> tf.keras.models.Model:

    if print_model_info:
        print_marquee("Model Summary")                                
        trained_model.summary()

    batch_size = fit_params.get("batch_size", BATCH_SIZE)

    print_marquee("Validation Dataset Confusion Matrix")
    
    val_gen = get_val_generator(fit_params["validation_data"],
                                preproc_func,
                                batch_size=batch_size)

    val_model_predictions = trained_model.predict(val_gen,
                                                  #steps=val_gen.n // (val_gen.batch_size + 1)
                                                 )
    
    pred_report = print_dataset_prediction_report(val_model_predictions,
                                                  val_gen.classes,
                                                  get_class_names(val_gen),
                                                  print_cm_as_percentages=True)

    if print_loss_graphs:
        
        print_marquee("Train/Val Accuracy and Loss graphs")

        # If using early stopping, it might be the case that we used less epochs than
        #  requested
        subtitle = f"Epochs: {len(trained_model_history.history['accuracy'])}"
        
        graph_loss_accuracy(trained_model_history,
                            subtitle=subtitle)
    
    return val_model_predictions, pred_report

# Helper functions for reporting

def print_dataset_prediction_report(y_pred,
                                    y_real,
                                    labels: list=None,
                                    print_cm_as_percentages: bool=False):
    
    # y_pred = np.round(y_pred).astype(int)
    y_pred = np.argmax(y_pred, axis=1)

    print_marquee("Classification Report")
    class_report_str = classification_report(y_real,
                                             y_pred,
                                             target_names=labels,
                                             digits=4)
    class_report_dict = classification_report(y_real,
                                              y_pred,
                                              target_names=labels,
                                              output_dict=True)
    print(class_report_str)
    
    print_marquee("Confusion Matrix")
    cm_data = confusion_matrix(y_real,y_pred)
    cm_display = ConfusionMatrixDisplay(confusion_matrix=cm_data,
                                        display_labels=labels)
    cm_display.plot(xticks_rotation="vertical")
    plt.show()

    if print_cm_as_percentages:
        print_marquee("Confusion Matrix as percentages")
        cm_data_perc = np.round(cm_data/np.sum(cm_data), 2)
        cm_display = ConfusionMatrixDisplay(confusion_matrix=cm_data_perc,
                                            display_labels=labels)
        cm_display.plot(xticks_rotation="vertical")
        plt.show()
    
    return class_report_dict


def graph_loss_accuracy(h_model,
                        subtitle: str = ""):
    
    num_records = len(h_model.history["accuracy"])
    
    plt.style.use("ggplot")
    plt.figure()
    plt.plot(np.arange(0,num_records),h_model.history["accuracy"],label="train_acc")
    plt.plot(np.arange(0,num_records),h_model.history["val_accuracy"],label="val_acc")
    plt.title("Training and Validation Accuracy" + f" ({subtitle})" if subtitle else "")
    plt.xlabel("Epoch")
    plt.ylabel("Accuracy")
    plt.legend()

    plt.figure()
    plt.plot(np.arange(0,num_records),h_model.history["loss"],label="train_loss") 
    plt.plot(np.arange(0,num_records),h_model.history["val_loss"],label="val_loss")
    plt.title("Training and Validation Loss" + f" ({subtitle})" if subtitle else "")
    plt.xlabel("Epoch")
    plt.ylabel("Loss")
    plt.legend()

    plt.show()

def create_model_checkpoint(filepath: str) -> tf.keras.callbacks.ModelCheckpoint:

    return tf.keras.callbacks.ModelCheckpoint(
            filepath,
            monitor = 'val_accuracy',
            verbose = 1,
            save_best_only = True,
            save_weights_only = False,
            mode = 'auto',
            # save_freq = 1,
        )

def calculate_cv_mean_metrics(cv_val_predictions):
    print_marquee(f"Mean metrics across {len(cv_val_predictions)} folds")
    display(pd.json_normalize(cv_val_predictions).mean().to_frame())

VGG16

%%time

#kf = KFold(n_splits=CV_FOLDS,
#           random_state=RANDOM_SEED,
#           shuffle=True)
kf = ShuffleSplit(n_splits=CV_FOLDS,
                  test_size=VAL_SIZE,
                  random_state=RANDOM_SEED)

split = 1

vgg16_cv_val_pred = []

for train_index, val_index in kf.split(balanced_training_data):

    vgg16_tf_model = tf.keras.applications.vgg16.VGG16(
        include_top=True,
        weights=None,
        pooling="avg",
        classes=NUM_CLASSES,
        classifier_activation='softmax',
    )

    # Add an optimizer
    vgg16_tf_model.compile(optimizer=tf.optimizers.Adam(learning_rate=0.003),
                        loss='categorical_crossentropy',
                        metrics=['accuracy'])
    #vgg16_tf_model.compile(optimizer=tf.optimizers.SGD(learning_rate=0.003, momentum=0.9),
    #                        loss='categorical_crossentropy',
    #                        metrics=['accuracy'])

    # Add a progress bar and save checkpoints
    vgg16_callbacks = [
        create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"vgg16_{TRAINING_RUN_ID}_{split}.h5")),
        tf.keras.callbacks.ProgbarLogger(
            count_mode = 'steps',
            stateful_metrics = None
        ),
        tf.keras.callbacks.EarlyStopping(
            monitor='val_accuracy',
            min_delta=0,
            patience=40,
            verbose=1,
            mode='auto',
            restore_best_weights=True
        )
    ]
        
    training_split_data = balanced_training_data.iloc[train_index]
    val_split_data = balanced_training_data.iloc[val_index]

    fit_params = {
        "x": training_split_data,
        "epochs": 105,
        "callbacks": vgg16_callbacks,
        "validation_data": val_split_data,
        "steps_per_epoch": 128,
        "validation_steps": 10,
    }

    preproc_func = tf.keras.applications.vgg16.preprocess_input

    with tf.device(TRAINING_DEVICE_NAME):

        vgg16_training_history = train_model(vgg16_tf_model,
                                             fit_params=fit_params,
                                             preproc_func=preproc_func)
        
        _, pred_report = evaluate_model(vgg16_tf_model,
                                        vgg16_training_history,
                                        fit_params=fit_params,
                                        preproc_func=preproc_func)
        
        vgg16_cv_val_pred.append(pred_report)
    
    split += 1

calculate_cv_mean_metrics(vgg16_cv_val_pred)

***************************
* Started at 10.734375... *
***************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
128/128 [==============================] - ETA: 0s - loss: 409.0229 - accuracy: 0.0767
Epoch 1: val_accuracy improved from -inf to 0.05625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 39s 234ms/step - loss: 409.0229 - accuracy: 0.0767 - val_loss: 2.7961 - val_accuracy: 0.0562
Epoch 2/105
128/128 [==============================] - ETA: 0s - loss: 2.5968 - accuracy: 0.0898
Epoch 2: val_accuracy improved from 0.05625 to 0.06250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 32s 249ms/step - loss: 2.5968 - accuracy: 0.0898 - val_loss: 2.4935 - val_accuracy: 0.0625
Epoch 3/105
128/128 [==============================] - ETA: 0s - loss: 2.4866 - accuracy: 0.0864
Epoch 3: val_accuracy improved from 0.06250 to 0.08750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 28s 218ms/step - loss: 2.4866 - accuracy: 0.0864 - val_loss: 2.4908 - val_accuracy: 0.0875
Epoch 4/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0806
Epoch 4: val_accuracy did not improve from 0.08750
128/128 [==============================] - 13s 102ms/step - loss: 2.4859 - accuracy: 0.0806 - val_loss: 2.4873 - val_accuracy: 0.0875
Epoch 5/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0781
Epoch 5: val_accuracy improved from 0.08750 to 0.10000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 35s 276ms/step - loss: 2.4860 - accuracy: 0.0781 - val_loss: 2.4851 - val_accuracy: 0.1000
Epoch 6/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0771
Epoch 6: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4852 - accuracy: 0.0771 - val_loss: 2.4871 - val_accuracy: 0.1000
Epoch 7/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0796
Epoch 7: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0796 - val_loss: 2.4843 - val_accuracy: 0.0625
Epoch 8/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0830
Epoch 8: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4848 - accuracy: 0.0830 - val_loss: 2.4841 - val_accuracy: 0.0875
Epoch 9/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0825
Epoch 9: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0825 - val_loss: 2.4842 - val_accuracy: 0.1000
Epoch 10/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0786
Epoch 10: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4855 - accuracy: 0.0786 - val_loss: 2.4882 - val_accuracy: 0.0500
Epoch 11/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0806
Epoch 11: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4864 - accuracy: 0.0806 - val_loss: 2.4875 - val_accuracy: 0.0625
Epoch 12/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0767
Epoch 12: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0767 - val_loss: 2.4860 - val_accuracy: 0.0688
Epoch 13/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0781
Epoch 13: val_accuracy improved from 0.10000 to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 27s 211ms/step - loss: 2.4853 - accuracy: 0.0781 - val_loss: 2.4850 - val_accuracy: 0.1063
Epoch 14/105
128/128 [==============================] - ETA: 0s - loss: 2.4849 - accuracy: 0.0830
Epoch 14: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4849 - accuracy: 0.0830 - val_loss: 2.4836 - val_accuracy: 0.1063
Epoch 15/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0801
Epoch 15: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0801 - val_loss: 2.4856 - val_accuracy: 0.1063
Epoch 16/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0806
Epoch 16: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4865 - accuracy: 0.0806 - val_loss: 2.4862 - val_accuracy: 0.0625
Epoch 17/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0859
Epoch 17: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0859 - val_loss: 2.4884 - val_accuracy: 0.0875
Epoch 18/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0933
Epoch 18: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4848 - accuracy: 0.0933 - val_loss: 2.4852 - val_accuracy: 0.1000
Epoch 19/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0830
Epoch 19: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4854 - accuracy: 0.0830 - val_loss: 2.4862 - val_accuracy: 0.1063
Epoch 20/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0845
Epoch 20: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4847 - accuracy: 0.0845 - val_loss: 2.4839 - val_accuracy: 0.1000
Epoch 21/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0874
Epoch 21: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0874 - val_loss: 2.4861 - val_accuracy: 0.1000
Epoch 22/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0835
Epoch 22: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0835 - val_loss: 2.4837 - val_accuracy: 0.1063
Epoch 23/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0874
Epoch 23: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0874 - val_loss: 2.4832 - val_accuracy: 0.1063
Epoch 24/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0752
Epoch 24: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0752 - val_loss: 2.4864 - val_accuracy: 0.0562
Epoch 25/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0762
Epoch 25: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0762 - val_loss: 2.4860 - val_accuracy: 0.1000
Epoch 26/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0801
Epoch 26: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0801 - val_loss: 2.4842 - val_accuracy: 0.0812
Epoch 27/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0776
Epoch 27: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4847 - accuracy: 0.0776 - val_loss: 2.4870 - val_accuracy: 0.0562
Epoch 28/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0830
Epoch 28: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4852 - accuracy: 0.0830 - val_loss: 2.4872 - val_accuracy: 0.0875
Epoch 29/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0879
Epoch 29: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0879 - val_loss: 2.4848 - val_accuracy: 0.0875
Epoch 30/105
128/128 [==============================] - ETA: 0s - loss: 2.4843 - accuracy: 0.0806
Epoch 30: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4843 - accuracy: 0.0806 - val_loss: 2.4878 - val_accuracy: 0.0688
Epoch 31/105
128/128 [==============================] - ETA: 0s - loss: 2.4849 - accuracy: 0.0791
Epoch 31: val_accuracy improved from 0.10625 to 0.11875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_1.h5
128/128 [==============================] - 34s 268ms/step - loss: 2.4849 - accuracy: 0.0791 - val_loss: 2.4867 - val_accuracy: 0.1187
Epoch 32/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0913
Epoch 32: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 102ms/step - loss: 2.4852 - accuracy: 0.0913 - val_loss: 2.4850 - val_accuracy: 0.0625
Epoch 33/105
128/128 [==============================] - ETA: 0s - loss: 2.4866 - accuracy: 0.0718
Epoch 33: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4866 - accuracy: 0.0718 - val_loss: 2.4848 - val_accuracy: 0.0562
Epoch 34/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0762
Epoch 34: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0762 - val_loss: 2.4841 - val_accuracy: 0.1187
Epoch 35/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0835
Epoch 35: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0835 - val_loss: 2.4865 - val_accuracy: 0.1187
Epoch 36/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0845
Epoch 36: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0845 - val_loss: 2.4863 - val_accuracy: 0.1187
Epoch 37/105
128/128 [==============================] - ETA: 0s - loss: 2.4830 - accuracy: 0.0952
Epoch 37: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4830 - accuracy: 0.0952 - val_loss: 2.4895 - val_accuracy: 0.0625
Epoch 38/105
128/128 [==============================] - ETA: 0s - loss: 2.4839 - accuracy: 0.0957
Epoch 38: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4839 - accuracy: 0.0957 - val_loss: 2.4904 - val_accuracy: 0.0625
Epoch 39/105
128/128 [==============================] - ETA: 0s - loss: 2.4841 - accuracy: 0.0928
Epoch 39: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4841 - accuracy: 0.0928 - val_loss: 2.4901 - val_accuracy: 0.0625
Epoch 40/105
128/128 [==============================] - ETA: 0s - loss: 2.4866 - accuracy: 0.0840
Epoch 40: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4866 - accuracy: 0.0840 - val_loss: 2.4857 - val_accuracy: 0.0625
Epoch 41/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0933
Epoch 41: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0933 - val_loss: 2.4905 - val_accuracy: 0.0625
Epoch 42/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0806
Epoch 42: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0806 - val_loss: 2.4891 - val_accuracy: 0.0625
Epoch 43/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0742
Epoch 43: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4855 - accuracy: 0.0742 - val_loss: 2.4897 - val_accuracy: 0.0812
Epoch 44/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0757
Epoch 44: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4865 - accuracy: 0.0757 - val_loss: 2.4896 - val_accuracy: 0.0500
Epoch 45/105
128/128 [==============================] - ETA: 0s - loss: 2.4861 - accuracy: 0.0708
Epoch 45: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4861 - accuracy: 0.0708 - val_loss: 2.4869 - val_accuracy: 0.1063
Epoch 46/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0928
Epoch 46: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0928 - val_loss: 2.4876 - val_accuracy: 0.0688
Epoch 47/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0762
Epoch 47: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4862 - accuracy: 0.0762 - val_loss: 2.4871 - val_accuracy: 0.0500
Epoch 48/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0781
Epoch 48: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0781 - val_loss: 2.4865 - val_accuracy: 0.0625
Epoch 49/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0820
Epoch 49: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0820 - val_loss: 2.4828 - val_accuracy: 0.0625
Epoch 50/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0952
Epoch 50: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0952 - val_loss: 2.4865 - val_accuracy: 0.0625
Epoch 51/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0913
Epoch 51: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 102ms/step - loss: 2.4851 - accuracy: 0.0913 - val_loss: 2.4912 - val_accuracy: 0.0500
Epoch 52/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0879
Epoch 52: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0879 - val_loss: 2.4917 - val_accuracy: 0.0625
Epoch 53/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0933
Epoch 53: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4851 - accuracy: 0.0933 - val_loss: 2.4888 - val_accuracy: 0.0625
Epoch 54/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0859
Epoch 54: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4864 - accuracy: 0.0859 - val_loss: 2.4885 - val_accuracy: 0.0625
Epoch 55/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0879
Epoch 55: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4844 - accuracy: 0.0879 - val_loss: 2.4895 - val_accuracy: 0.0625
Epoch 56/105
128/128 [==============================] - ETA: 0s - loss: 2.4861 - accuracy: 0.0913
Epoch 56: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4861 - accuracy: 0.0913 - val_loss: 2.4865 - val_accuracy: 0.0625
Epoch 57/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0674
Epoch 57: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4860 - accuracy: 0.0674 - val_loss: 2.4864 - val_accuracy: 0.0812
Epoch 58/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0742
Epoch 58: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4860 - accuracy: 0.0742 - val_loss: 2.4860 - val_accuracy: 0.1000
Epoch 59/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0767
Epoch 59: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0767 - val_loss: 2.4875 - val_accuracy: 0.0500
Epoch 60/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0864
Epoch 60: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 102ms/step - loss: 2.4847 - accuracy: 0.0864 - val_loss: 2.4876 - val_accuracy: 0.0500
Epoch 61/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0762
Epoch 61: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0762 - val_loss: 2.4861 - val_accuracy: 0.1000
Epoch 62/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0781
Epoch 62: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0781 - val_loss: 2.4859 - val_accuracy: 0.1000
Epoch 63/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0742
Epoch 63: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4865 - accuracy: 0.0742 - val_loss: 2.4879 - val_accuracy: 0.0500
Epoch 64/105
128/128 [==============================] - ETA: 0s - loss: 2.4850 - accuracy: 0.0835
Epoch 64: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4850 - accuracy: 0.0835 - val_loss: 2.4861 - val_accuracy: 0.0875
Epoch 65/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0825
Epoch 65: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4862 - accuracy: 0.0825 - val_loss: 2.4869 - val_accuracy: 0.0625
Epoch 66/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0972
Epoch 66: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4848 - accuracy: 0.0972 - val_loss: 2.4862 - val_accuracy: 0.0625
Epoch 67/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0874
Epoch 67: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0874 - val_loss: 2.4869 - val_accuracy: 0.0625
Epoch 68/105
128/128 [==============================] - ETA: 0s - loss: 2.4850 - accuracy: 0.0869
Epoch 68: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4850 - accuracy: 0.0869 - val_loss: 2.4880 - val_accuracy: 0.0688
Epoch 69/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0781
Epoch 69: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4862 - accuracy: 0.0781 - val_loss: 2.4870 - val_accuracy: 0.0625
Epoch 70/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0791
Epoch 70: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0791 - val_loss: 2.4853 - val_accuracy: 0.0625
Epoch 71/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0684
Epoch 71: val_accuracy did not improve from 0.11875
128/128 [==============================] - 13s 103ms/step - loss: 2.4862 - accuracy: 0.0684 - val_loss: 2.4858 - val_accuracy: 0.1063
Restoring model weights from the end of the best epoch: 31.
Epoch 71: early stopping


 ********* Training time: 2052.109375 s.


*****************
* Model Summary *
*****************
Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 predictions (Dense)         (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 134,309,708
Non-trainable params: 0
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 27ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.0000    0.0000    0.0000       305
anarrhichomenum     0.0000    0.0000    0.0000       322
   brevantherum     0.0000    0.0000    0.0000       266
      dulcamara     0.0000    0.0000    0.0000       311
   herposolanum     0.0000    0.0000    0.0000       279
     holophylla     0.0000    0.0000    0.0000       282
     lasiocarpa     0.0869    1.0000    0.1600       313
      melongena     0.0000    0.0000    0.0000       314
    micracantha     0.0000    0.0000    0.0000       299
         petota     0.0000    0.0000    0.0000       300
        solanum     0.0000    0.0000    0.0000       304
          torva     0.0000    0.0000    0.0000       305

       accuracy                         0.0869      3600
      macro avg     0.0072    0.0833    0.0133      3600
   weighted avg     0.0076    0.0869    0.0139      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

****************************
* Started at 2074.59375... *
****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
128/128 [==============================] - ETA: 0s - loss: 140.9845 - accuracy: 0.0869
Epoch 1: val_accuracy improved from -inf to 0.08125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_2.h5
128/128 [==============================] - 27s 207ms/step - loss: 140.9845 - accuracy: 0.0869 - val_loss: 2.4867 - val_accuracy: 0.0812
Epoch 2/105
128/128 [==============================] - ETA: 0s - loss: 2.4871 - accuracy: 0.0884
Epoch 2: val_accuracy did not improve from 0.08125
128/128 [==============================] - 13s 102ms/step - loss: 2.4871 - accuracy: 0.0884 - val_loss: 2.4844 - val_accuracy: 0.0812
Epoch 3/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0811
Epoch 3: val_accuracy improved from 0.08125 to 0.10000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_2.h5
128/128 [==============================] - 27s 212ms/step - loss: 2.4860 - accuracy: 0.0811 - val_loss: 2.4847 - val_accuracy: 0.1000
Epoch 4/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0801
Epoch 4: val_accuracy did not improve from 0.10000
128/128 [==============================] - 14s 102ms/step - loss: 2.4858 - accuracy: 0.0801 - val_loss: 2.4855 - val_accuracy: 0.0688
Epoch 5/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0801
Epoch 5: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0801 - val_loss: 2.4858 - val_accuracy: 0.0812
Epoch 6/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0898
Epoch 6: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4853 - accuracy: 0.0898 - val_loss: 2.4859 - val_accuracy: 0.0812
Epoch 7/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0825
Epoch 7: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0825 - val_loss: 2.4861 - val_accuracy: 0.0812
Epoch 8/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0840
Epoch 8: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0840 - val_loss: 2.4867 - val_accuracy: 0.0812
Epoch 9/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0874
Epoch 9: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0874 - val_loss: 2.4854 - val_accuracy: 0.0812
Epoch 10/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0845
Epoch 10: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0845 - val_loss: 2.4842 - val_accuracy: 0.0812
Epoch 11/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0806
Epoch 11: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4844 - accuracy: 0.0806 - val_loss: 2.4842 - val_accuracy: 0.0812
Epoch 12/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0864
Epoch 12: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0864 - val_loss: 2.4837 - val_accuracy: 0.0812
Epoch 13/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0835
Epoch 13: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4848 - accuracy: 0.0835 - val_loss: 2.4836 - val_accuracy: 0.0812
Epoch 14/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0845
Epoch 14: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0845 - val_loss: 2.4832 - val_accuracy: 0.0812
Epoch 15/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0889
Epoch 15: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4852 - accuracy: 0.0889 - val_loss: 2.4849 - val_accuracy: 0.0812
Epoch 16/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0796
Epoch 16: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0796 - val_loss: 2.4856 - val_accuracy: 0.0750
Epoch 17/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0791
Epoch 17: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0791 - val_loss: 2.4851 - val_accuracy: 0.0812
Epoch 18/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0757
Epoch 18: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0757 - val_loss: 2.4854 - val_accuracy: 0.0625
Epoch 19/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0747
Epoch 19: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4851 - accuracy: 0.0747 - val_loss: 2.4855 - val_accuracy: 0.0812
Epoch 20/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0801
Epoch 20: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0801 - val_loss: 2.4864 - val_accuracy: 0.0688
Epoch 21/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0864
Epoch 21: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0864 - val_loss: 2.4886 - val_accuracy: 0.0688
Epoch 22/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0825
Epoch 22: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0825 - val_loss: 2.4874 - val_accuracy: 0.0625
Epoch 23/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0786
Epoch 23: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0786 - val_loss: 2.4868 - val_accuracy: 0.0750
Epoch 24/105
128/128 [==============================] - ETA: 0s - loss: 2.4845 - accuracy: 0.0776
Epoch 24: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4845 - accuracy: 0.0776 - val_loss: 2.4860 - val_accuracy: 0.0812
Epoch 25/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0835
Epoch 25: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4865 - accuracy: 0.0835 - val_loss: 2.4846 - val_accuracy: 0.1000
Epoch 26/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0781
Epoch 26: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4865 - accuracy: 0.0781 - val_loss: 2.4849 - val_accuracy: 0.1000
Epoch 27/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0825
Epoch 27: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4855 - accuracy: 0.0825 - val_loss: 2.4857 - val_accuracy: 0.0750
Epoch 28/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0908
Epoch 28: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4854 - accuracy: 0.0908 - val_loss: 2.4867 - val_accuracy: 0.0812
Epoch 29/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0776
Epoch 29: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4854 - accuracy: 0.0776 - val_loss: 2.4876 - val_accuracy: 0.0812
Epoch 30/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0928
Epoch 30: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4848 - accuracy: 0.0928 - val_loss: 2.4871 - val_accuracy: 0.0812
Epoch 31/105
128/128 [==============================] - ETA: 0s - loss: 2.4843 - accuracy: 0.0903
Epoch 31: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4843 - accuracy: 0.0903 - val_loss: 2.4880 - val_accuracy: 0.0750
Epoch 32/105
128/128 [==============================] - ETA: 0s - loss: 2.4870 - accuracy: 0.0825
Epoch 32: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 102ms/step - loss: 2.4870 - accuracy: 0.0825 - val_loss: 2.4868 - val_accuracy: 0.0750
Epoch 33/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0835
Epoch 33: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0835 - val_loss: 2.4868 - val_accuracy: 0.0812
Epoch 34/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0869
Epoch 34: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0869 - val_loss: 2.4882 - val_accuracy: 0.0812
Epoch 35/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0859
Epoch 35: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4865 - accuracy: 0.0859 - val_loss: 2.4863 - val_accuracy: 0.0812
Epoch 36/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0825
Epoch 36: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 104ms/step - loss: 2.4856 - accuracy: 0.0825 - val_loss: 2.4862 - val_accuracy: 0.0812
Epoch 37/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0825
Epoch 37: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4848 - accuracy: 0.0825 - val_loss: 2.4885 - val_accuracy: 0.0812
Epoch 38/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0820
Epoch 38: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4853 - accuracy: 0.0820 - val_loss: 2.4870 - val_accuracy: 0.0812
Epoch 39/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0757
Epoch 39: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0757 - val_loss: 2.4869 - val_accuracy: 0.0812
Epoch 40/105
128/128 [==============================] - ETA: 0s - loss: 2.4861 - accuracy: 0.0747
Epoch 40: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4861 - accuracy: 0.0747 - val_loss: 2.4853 - val_accuracy: 0.0688
Epoch 41/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0815
Epoch 41: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4858 - accuracy: 0.0815 - val_loss: 2.4836 - val_accuracy: 0.1000
Epoch 42/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0898
Epoch 42: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4847 - accuracy: 0.0898 - val_loss: 2.4836 - val_accuracy: 0.0812
Epoch 43/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0771
Epoch 43: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0771 - val_loss: 2.4861 - val_accuracy: 0.0812
Restoring model weights from the end of the best epoch: 3.
Epoch 43: early stopping


 ********* Training time: 1237.046875 s.


*****************
* Model Summary *
*****************
Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 predictions (Dense)         (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 134,309,708
Non-trainable params: 0
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 25ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.0000    0.0000    0.0000       277
anarrhichomenum     0.0000    0.0000    0.0000       315
   brevantherum     0.0000    0.0000    0.0000       307
      dulcamara     0.0825    1.0000    0.1524       297
   herposolanum     0.0000    0.0000    0.0000       288
     holophylla     0.0000    0.0000    0.0000       286
     lasiocarpa     0.0000    0.0000    0.0000       314
      melongena     0.0000    0.0000    0.0000       291
    micracantha     0.0000    0.0000    0.0000       292
         petota     0.0000    0.0000    0.0000       328
        solanum     0.0000    0.0000    0.0000       304
          torva     0.0000    0.0000    0.0000       301

       accuracy                         0.0825      3600
      macro avg     0.0069    0.0833    0.0127      3600
   weighted avg     0.0068    0.0825    0.0126      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*****************************
* Started at 3323.390625... *
*****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
128/128 [==============================] - ETA: 0s - loss: 76.3238 - accuracy: 0.0898
Epoch 1: val_accuracy improved from -inf to 0.05000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_3.h5
128/128 [==============================] - 28s 213ms/step - loss: 76.3238 - accuracy: 0.0898 - val_loss: 2.4914 - val_accuracy: 0.0500
Epoch 2/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0752
Epoch 2: val_accuracy improved from 0.05000 to 0.07500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_3.h5
128/128 [==============================] - 35s 273ms/step - loss: 2.4865 - accuracy: 0.0752 - val_loss: 2.4854 - val_accuracy: 0.0750
Epoch 3/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0845
Epoch 3: val_accuracy improved from 0.07500 to 0.11250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_3.h5
128/128 [==============================] - 33s 256ms/step - loss: 2.4844 - accuracy: 0.0845 - val_loss: 2.4889 - val_accuracy: 0.1125
Epoch 4/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0786
Epoch 4: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4857 - accuracy: 0.0786 - val_loss: 2.4866 - val_accuracy: 0.0625
Epoch 5/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0747
Epoch 5: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4854 - accuracy: 0.0747 - val_loss: 2.4847 - val_accuracy: 0.1000
Epoch 6/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0752
Epoch 6: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4857 - accuracy: 0.0752 - val_loss: 2.4835 - val_accuracy: 0.0875
Epoch 7/105
128/128 [==============================] - ETA: 0s - loss: 2.4845 - accuracy: 0.0869
Epoch 7: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4845 - accuracy: 0.0869 - val_loss: 2.4791 - val_accuracy: 0.1000
Epoch 8/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0747
Epoch 8: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4864 - accuracy: 0.0747 - val_loss: 2.4824 - val_accuracy: 0.0750
Epoch 9/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0874
Epoch 9: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4856 - accuracy: 0.0874 - val_loss: 2.4872 - val_accuracy: 0.0688
Epoch 10/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0840
Epoch 10: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4859 - accuracy: 0.0840 - val_loss: 2.4873 - val_accuracy: 0.0688
Epoch 11/105
128/128 [==============================] - ETA: 0s - loss: 2.4843 - accuracy: 0.0952
Epoch 11: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4843 - accuracy: 0.0952 - val_loss: 2.4862 - val_accuracy: 0.0812
Epoch 12/105
128/128 [==============================] - ETA: 0s - loss: 2.4853 - accuracy: 0.0879
Epoch 12: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4853 - accuracy: 0.0879 - val_loss: 2.4894 - val_accuracy: 0.0812
Epoch 13/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0791
Epoch 13: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4856 - accuracy: 0.0791 - val_loss: 2.4853 - val_accuracy: 0.1125
Epoch 14/105
128/128 [==============================] - ETA: 0s - loss: 2.4861 - accuracy: 0.0776
Epoch 14: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4861 - accuracy: 0.0776 - val_loss: 2.4877 - val_accuracy: 0.1000
Epoch 15/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0894
Epoch 15: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4854 - accuracy: 0.0894 - val_loss: 2.4882 - val_accuracy: 0.0688
Epoch 16/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0742
Epoch 16: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4860 - accuracy: 0.0742 - val_loss: 2.4895 - val_accuracy: 0.0625
Epoch 17/105
128/128 [==============================] - ETA: 0s - loss: 2.4849 - accuracy: 0.0859
Epoch 17: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4849 - accuracy: 0.0859 - val_loss: 2.4913 - val_accuracy: 0.0750
Epoch 18/105
128/128 [==============================] - ETA: 0s - loss: 2.4846 - accuracy: 0.0801
Epoch 18: val_accuracy did not improve from 0.11250
128/128 [==============================] - 14s 106ms/step - loss: 2.4846 - accuracy: 0.0801 - val_loss: 2.4935 - val_accuracy: 0.0500
Epoch 19/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0781
Epoch 19: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4862 - accuracy: 0.0781 - val_loss: 2.4894 - val_accuracy: 0.0750
Epoch 20/105
128/128 [==============================] - ETA: 0s - loss: 2.4850 - accuracy: 0.0845
Epoch 20: val_accuracy did not improve from 0.11250
128/128 [==============================] - 14s 111ms/step - loss: 2.4850 - accuracy: 0.0845 - val_loss: 2.4888 - val_accuracy: 0.1000
Epoch 21/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0820
Epoch 21: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 101ms/step - loss: 2.4864 - accuracy: 0.0820 - val_loss: 2.4869 - val_accuracy: 0.0812
Epoch 22/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0742
Epoch 22: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4860 - accuracy: 0.0742 - val_loss: 2.4883 - val_accuracy: 0.0688
Epoch 23/105
128/128 [==============================] - ETA: 0s - loss: 2.4849 - accuracy: 0.0903
Epoch 23: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4849 - accuracy: 0.0903 - val_loss: 2.4889 - val_accuracy: 0.0688
Epoch 24/105
128/128 [==============================] - ETA: 0s - loss: 2.4840 - accuracy: 0.0894
Epoch 24: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4840 - accuracy: 0.0894 - val_loss: 2.4881 - val_accuracy: 0.1000
Epoch 25/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0874
Epoch 25: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4865 - accuracy: 0.0874 - val_loss: 2.4868 - val_accuracy: 0.1000
Epoch 26/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0723
Epoch 26: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4860 - accuracy: 0.0723 - val_loss: 2.4869 - val_accuracy: 0.1000
Epoch 27/105
128/128 [==============================] - ETA: 0s - loss: 2.4843 - accuracy: 0.0938
Epoch 27: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 105ms/step - loss: 2.4843 - accuracy: 0.0938 - val_loss: 2.4847 - val_accuracy: 0.1125
Epoch 28/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0781
Epoch 28: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4859 - accuracy: 0.0781 - val_loss: 2.4822 - val_accuracy: 0.1125
Epoch 29/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0962
Epoch 29: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0962 - val_loss: 2.4835 - val_accuracy: 0.1000
Epoch 30/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0815
Epoch 30: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0815 - val_loss: 2.4858 - val_accuracy: 0.1000
Epoch 31/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0732
Epoch 31: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4860 - accuracy: 0.0732 - val_loss: 2.4895 - val_accuracy: 0.0688
Epoch 32/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0869
Epoch 32: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4864 - accuracy: 0.0869 - val_loss: 2.4846 - val_accuracy: 0.0875
Epoch 33/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0869
Epoch 33: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4855 - accuracy: 0.0869 - val_loss: 2.4850 - val_accuracy: 0.0688
Epoch 34/105
128/128 [==============================] - ETA: 0s - loss: 2.4850 - accuracy: 0.0864
Epoch 34: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4850 - accuracy: 0.0864 - val_loss: 2.4897 - val_accuracy: 0.0688
Epoch 35/105
128/128 [==============================] - ETA: 0s - loss: 2.4841 - accuracy: 0.0874
Epoch 35: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4841 - accuracy: 0.0874 - val_loss: 2.4934 - val_accuracy: 0.0625
Epoch 36/105
128/128 [==============================] - ETA: 0s - loss: 2.4849 - accuracy: 0.0850
Epoch 36: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4849 - accuracy: 0.0850 - val_loss: 2.4923 - val_accuracy: 0.0688
Epoch 37/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0923
Epoch 37: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4847 - accuracy: 0.0923 - val_loss: 2.4904 - val_accuracy: 0.0688
Epoch 38/105
128/128 [==============================] - ETA: 0s - loss: 2.4865 - accuracy: 0.0801
Epoch 38: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4865 - accuracy: 0.0801 - val_loss: 2.4874 - val_accuracy: 0.0688
Epoch 39/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0884
Epoch 39: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4854 - accuracy: 0.0884 - val_loss: 2.4877 - val_accuracy: 0.0688
Epoch 40/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0854
Epoch 40: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0854 - val_loss: 2.4878 - val_accuracy: 0.0688
Epoch 41/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0757
Epoch 41: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 102ms/step - loss: 2.4855 - accuracy: 0.0757 - val_loss: 2.4883 - val_accuracy: 0.0625
Epoch 42/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0781
Epoch 42: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0781 - val_loss: 2.4865 - val_accuracy: 0.0812
Epoch 43/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0767
Epoch 43: val_accuracy did not improve from 0.11250
128/128 [==============================] - 13s 103ms/step - loss: 2.4860 - accuracy: 0.0767 - val_loss: 2.4852 - val_accuracy: 0.0812
Restoring model weights from the end of the best epoch: 3.
Epoch 43: early stopping


 ********* Training time: 1226.921875 s.


*****************
* Model Summary *
*****************
Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_3 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 predictions (Dense)         (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 134,309,708
Non-trainable params: 0
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 26ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.0000    0.0000    0.0000       280
anarrhichomenum     0.0000    0.0000    0.0000       319
   brevantherum     0.0000    0.0000    0.0000       267
      dulcamara     0.0000    0.0000    0.0000       315
   herposolanum     0.0000    0.0000    0.0000       300
     holophylla     0.0000    0.0000    0.0000       335
     lasiocarpa     0.0858    1.0000    0.1581       309
      melongena     0.0000    0.0000    0.0000       285
    micracantha     0.0000    0.0000    0.0000       294
         petota     0.0000    0.0000    0.0000       301
        solanum     0.0000    0.0000    0.0000       299
          torva     0.0000    0.0000    0.0000       296

       accuracy                         0.0858      3600
      macro avg     0.0072    0.0833    0.0132      3600
   weighted avg     0.0074    0.0858    0.0136      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

****************************
* Started at 4561.78125... *
****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
128/128 [==============================] - ETA: 0s - loss: 201.7437 - accuracy: 0.0762
Epoch 1: val_accuracy improved from -inf to 0.06875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_4.h5
128/128 [==============================] - 31s 236ms/step - loss: 201.7437 - accuracy: 0.0762 - val_loss: 2.4879 - val_accuracy: 0.0688
Epoch 2/105
128/128 [==============================] - ETA: 0s - loss: 2.4928 - accuracy: 0.0889
Epoch 2: val_accuracy did not improve from 0.06875
128/128 [==============================] - 13s 101ms/step - loss: 2.4928 - accuracy: 0.0889 - val_loss: 2.4808 - val_accuracy: 0.0688
Epoch 3/105
128/128 [==============================] - ETA: 0s - loss: 2.4869 - accuracy: 0.0840
Epoch 3: val_accuracy improved from 0.06875 to 0.09375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_4.h5
128/128 [==============================] - 36s 285ms/step - loss: 2.4869 - accuracy: 0.0840 - val_loss: 2.4837 - val_accuracy: 0.0938
Epoch 4/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0820
Epoch 4: val_accuracy did not improve from 0.09375
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0820 - val_loss: 2.4830 - val_accuracy: 0.0938
Epoch 5/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0845
Epoch 5: val_accuracy did not improve from 0.09375
128/128 [==============================] - 13s 101ms/step - loss: 2.4848 - accuracy: 0.0845 - val_loss: 2.4827 - val_accuracy: 0.0938
Epoch 6/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0796
Epoch 6: val_accuracy improved from 0.09375 to 0.10000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_4.h5
128/128 [==============================] - 31s 240ms/step - loss: 2.4857 - accuracy: 0.0796 - val_loss: 2.4842 - val_accuracy: 0.1000
Epoch 7/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0801
Epoch 7: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 101ms/step - loss: 2.4855 - accuracy: 0.0801 - val_loss: 2.4877 - val_accuracy: 0.0625
Epoch 8/105
128/128 [==============================] - ETA: 0s - loss: 2.4860 - accuracy: 0.0928
Epoch 8: val_accuracy did not improve from 0.10000
128/128 [==============================] - 13s 101ms/step - loss: 2.4860 - accuracy: 0.0928 - val_loss: 2.4853 - val_accuracy: 0.0625
Epoch 9/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0806
Epoch 9: val_accuracy improved from 0.10000 to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_20230216171100_4.h5
128/128 [==============================] - 35s 276ms/step - loss: 2.4855 - accuracy: 0.0806 - val_loss: 2.4850 - val_accuracy: 0.1063
Epoch 10/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0806
Epoch 10: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4859 - accuracy: 0.0806 - val_loss: 2.4849 - val_accuracy: 0.1063
Epoch 11/105
128/128 [==============================] - ETA: 0s - loss: 2.4846 - accuracy: 0.0859
Epoch 11: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4846 - accuracy: 0.0859 - val_loss: 2.4857 - val_accuracy: 0.1063
Epoch 12/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0898
Epoch 12: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4855 - accuracy: 0.0898 - val_loss: 2.4863 - val_accuracy: 0.1063
Epoch 13/105
128/128 [==============================] - ETA: 0s - loss: 2.4847 - accuracy: 0.0820
Epoch 13: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4847 - accuracy: 0.0820 - val_loss: 2.4853 - val_accuracy: 0.0625
Epoch 14/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0850
Epoch 14: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4856 - accuracy: 0.0850 - val_loss: 2.4858 - val_accuracy: 0.1000
Epoch 15/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0825
Epoch 15: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4855 - accuracy: 0.0825 - val_loss: 2.4897 - val_accuracy: 0.0750
Epoch 16/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0796
Epoch 16: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4852 - accuracy: 0.0796 - val_loss: 2.4846 - val_accuracy: 0.1000
Epoch 17/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0854
Epoch 17: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4856 - accuracy: 0.0854 - val_loss: 2.4862 - val_accuracy: 0.0750
Epoch 18/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0825
Epoch 18: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0825 - val_loss: 2.4899 - val_accuracy: 0.0562
Epoch 19/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0815
Epoch 19: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4859 - accuracy: 0.0815 - val_loss: 2.4903 - val_accuracy: 0.0562
Epoch 20/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0771
Epoch 20: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4857 - accuracy: 0.0771 - val_loss: 2.4899 - val_accuracy: 0.0688
Epoch 21/105
128/128 [==============================] - ETA: 0s - loss: 2.4855 - accuracy: 0.0854
Epoch 21: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4855 - accuracy: 0.0854 - val_loss: 2.4891 - val_accuracy: 0.0562
Epoch 22/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0854
Epoch 22: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4864 - accuracy: 0.0854 - val_loss: 2.4873 - val_accuracy: 0.0562
Epoch 23/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0801
Epoch 23: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0801 - val_loss: 2.4868 - val_accuracy: 0.1000
Epoch 24/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0830
Epoch 24: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 104ms/step - loss: 2.4844 - accuracy: 0.0830 - val_loss: 2.4882 - val_accuracy: 0.1000
Epoch 25/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0859
Epoch 25: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0859 - val_loss: 2.4888 - val_accuracy: 0.0562
Epoch 26/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0874
Epoch 26: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0874 - val_loss: 2.4874 - val_accuracy: 0.0562
Epoch 27/105
128/128 [==============================] - ETA: 0s - loss: 2.4845 - accuracy: 0.0864
Epoch 27: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4845 - accuracy: 0.0864 - val_loss: 2.4880 - val_accuracy: 0.1063
Epoch 28/105
128/128 [==============================] - ETA: 0s - loss: 2.4864 - accuracy: 0.0815
Epoch 28: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4864 - accuracy: 0.0815 - val_loss: 2.4851 - val_accuracy: 0.1063
Epoch 29/105
128/128 [==============================] - ETA: 0s - loss: 2.4859 - accuracy: 0.0820
Epoch 29: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4859 - accuracy: 0.0820 - val_loss: 2.4854 - val_accuracy: 0.1000
Epoch 30/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0845
Epoch 30: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4851 - accuracy: 0.0845 - val_loss: 2.4855 - val_accuracy: 0.0625
Epoch 31/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0835
Epoch 31: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4844 - accuracy: 0.0835 - val_loss: 2.4891 - val_accuracy: 0.0562
Epoch 32/105
128/128 [==============================] - ETA: 0s - loss: 2.4835 - accuracy: 0.0903
Epoch 32: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4835 - accuracy: 0.0903 - val_loss: 2.4885 - val_accuracy: 0.1063
Epoch 33/105
128/128 [==============================] - ETA: 0s - loss: 2.4873 - accuracy: 0.0815
Epoch 33: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4873 - accuracy: 0.0815 - val_loss: 2.4854 - val_accuracy: 0.1063
Epoch 34/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0728
Epoch 34: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0728 - val_loss: 2.4841 - val_accuracy: 0.0562
Epoch 35/105
128/128 [==============================] - ETA: 0s - loss: 2.4871 - accuracy: 0.0767
Epoch 35: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4871 - accuracy: 0.0767 - val_loss: 2.4861 - val_accuracy: 0.0562
Epoch 36/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0879
Epoch 36: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0879 - val_loss: 2.4851 - val_accuracy: 0.0562
Epoch 37/105
128/128 [==============================] - ETA: 0s - loss: 2.4856 - accuracy: 0.0845
Epoch 37: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4856 - accuracy: 0.0845 - val_loss: 2.4841 - val_accuracy: 0.1000
Epoch 38/105
128/128 [==============================] - ETA: 0s - loss: 2.4846 - accuracy: 0.0928
Epoch 38: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4846 - accuracy: 0.0928 - val_loss: 2.4837 - val_accuracy: 0.1000
Epoch 39/105
128/128 [==============================] - ETA: 0s - loss: 2.4848 - accuracy: 0.0898
Epoch 39: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4848 - accuracy: 0.0898 - val_loss: 2.4833 - val_accuracy: 0.1000
Epoch 40/105
128/128 [==============================] - ETA: 0s - loss: 2.4844 - accuracy: 0.0845
Epoch 40: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4844 - accuracy: 0.0845 - val_loss: 2.4857 - val_accuracy: 0.1063
Epoch 41/105
128/128 [==============================] - ETA: 0s - loss: 2.4862 - accuracy: 0.0864
Epoch 41: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4862 - accuracy: 0.0864 - val_loss: 2.4858 - val_accuracy: 0.1000
Epoch 42/105
128/128 [==============================] - ETA: 0s - loss: 2.4857 - accuracy: 0.0767
Epoch 42: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4857 - accuracy: 0.0767 - val_loss: 2.4876 - val_accuracy: 0.1063
Epoch 43/105
128/128 [==============================] - ETA: 0s - loss: 2.4861 - accuracy: 0.0767
Epoch 43: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4861 - accuracy: 0.0767 - val_loss: 2.4874 - val_accuracy: 0.0562
Epoch 44/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0840
Epoch 44: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4851 - accuracy: 0.0840 - val_loss: 2.4874 - val_accuracy: 0.1063
Epoch 45/105
128/128 [==============================] - ETA: 0s - loss: 2.4851 - accuracy: 0.0811
Epoch 45: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4851 - accuracy: 0.0811 - val_loss: 2.4908 - val_accuracy: 0.0625
Epoch 46/105
128/128 [==============================] - ETA: 0s - loss: 2.4852 - accuracy: 0.0918
Epoch 46: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4852 - accuracy: 0.0918 - val_loss: 2.4897 - val_accuracy: 0.0625
Epoch 47/105
128/128 [==============================] - ETA: 0s - loss: 2.4854 - accuracy: 0.0806
Epoch 47: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4854 - accuracy: 0.0806 - val_loss: 2.4902 - val_accuracy: 0.0688
Epoch 48/105
128/128 [==============================] - ETA: 0s - loss: 2.4863 - accuracy: 0.0840
Epoch 48: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 103ms/step - loss: 2.4863 - accuracy: 0.0840 - val_loss: 2.4889 - val_accuracy: 0.0562
Epoch 49/105
128/128 [==============================] - ETA: 0s - loss: 2.4858 - accuracy: 0.0884
Epoch 49: val_accuracy did not improve from 0.10625
128/128 [==============================] - 13s 102ms/step - loss: 2.4858 - accuracy: 0.0884 - val_loss: 2.4853 - val_accuracy: 0.0562
Restoring model weights from the end of the best epoch: 9.
Epoch 49: early stopping


 ********* Training time: 1398.515625 s.


*****************
* Model Summary *
*****************
Model: "vgg16"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_4 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 predictions (Dense)         (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 134,309,708
Non-trainable params: 0
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 26ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.0000    0.0000    0.0000       331
anarrhichomenum     0.0000    0.0000    0.0000       319
   brevantherum     0.0000    0.0000    0.0000       312
      dulcamara     0.0000    0.0000    0.0000       259
   herposolanum     0.0000    0.0000    0.0000       273
     holophylla     0.0000    0.0000    0.0000       316
     lasiocarpa     0.0000    0.0000    0.0000       314
      melongena     0.0000    0.0000    0.0000       302
    micracantha     0.0875    1.0000    0.1609       315
         petota     0.0000    0.0000    0.0000       289
        solanum     0.0000    0.0000    0.0000       271
          torva     0.0000    0.0000    0.0000       299

       accuracy                         0.0875      3600
      macro avg     0.0073    0.0833    0.0134      3600
   weighted avg     0.0077    0.0875    0.0141      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*******************************
* Mean metrics across 4 folds *
*******************************

	0
accuracy	0.085694
acanthophora.precision	0.000000
acanthophora.recall	0.000000
acanthophora.f1-score	0.000000
acanthophora.support	298.250000
anarrhichomenum.precision	0.000000
anarrhichomenum.recall	0.000000
anarrhichomenum.f1-score	0.000000
anarrhichomenum.support	318.750000
brevantherum.precision	0.000000
brevantherum.recall	0.000000
brevantherum.f1-score	0.000000
brevantherum.support	288.000000
dulcamara.precision	0.020625
dulcamara.recall	0.250000
dulcamara.f1-score	0.038106
dulcamara.support	295.500000
herposolanum.precision	0.000000
herposolanum.recall	0.000000
herposolanum.f1-score	0.000000
herposolanum.support	285.000000
holophylla.precision	0.000000
holophylla.recall	0.000000
holophylla.f1-score	0.000000
holophylla.support	304.750000
lasiocarpa.precision	0.043194
lasiocarpa.recall	0.500000
lasiocarpa.f1-score	0.079519
lasiocarpa.support	312.500000
melongena.precision	0.000000
melongena.recall	0.000000
melongena.f1-score	0.000000
melongena.support	298.000000
micracantha.precision	0.021875
micracantha.recall	0.250000
micracantha.f1-score	0.040230
micracantha.support	300.000000
petota.precision	0.000000
petota.recall	0.000000
petota.f1-score	0.000000
petota.support	304.500000
solanum.precision	0.000000
solanum.recall	0.000000
solanum.f1-score	0.000000
solanum.support	294.500000
torva.precision	0.000000
torva.recall	0.000000
torva.f1-score	0.000000
torva.support	300.250000
macro avg.precision	0.007141
macro avg.recall	0.083333
macro avg.f1-score	0.013155
macro avg.support	3600.000000
weighted avg.precision	0.007347
weighted avg.recall	0.085694
weighted avg.f1-score	0.013534
weighted avg.support	3600.000000

CPU times: total: 1h 39min 21s
Wall time: 50min 45s

VGG16 with ImageNet Transfer Learning

%%time

kf = ShuffleSplit(n_splits=CV_FOLDS,
                  test_size=VAL_SIZE,
                  random_state=RANDOM_SEED)

split = 1

vgg16_tl_cv_val_pred = []

for train_index, val_index in kf.split(balanced_training_data):

    vgg16_tl_model = tf.keras.applications.vgg16.VGG16(
        include_top=True,
        weights='imagenet',
        pooling="avg",
        # classes=12,
        classifier_activation='softmax',
    )

    # Change the output prediction layer to support the 12 classes instead of the 1000 classes in ImageNet
    output = vgg16_tl_model.layers[-2].output
    predictions = tf.keras.layers.Dense(NUM_CLASSES,
                                        activation="softmax")(output)
    vgg16_tl_model = tf.keras.Model(inputs = vgg16_tl_model.input,
                                    outputs = predictions)

    # Freeze all layers but the FC and output
    for layer in (vgg16_tl_model.layers)[:19]:
        layer.trainable = False

    # Add an optimizer
    vgg16_tl_model.compile(optimizer=tf.optimizers.Adam(learning_rate=0.003),
                        loss='categorical_crossentropy',
                        metrics=['accuracy'])
    #vgg16_tl_model.compile(optimizer=tf.optimizers.SGD(learning_rate=0.0001, momentum=0.9),
    #                        loss='categorical_crossentropy',
    #                        metrics=['accuracy'])
        
    training_split_data = balanced_training_data.iloc[train_index]
    val_split_data = balanced_training_data.iloc[val_index]

    # Add a progress bar and save checkpoints
    vgg16_tl_callbacks = [
        create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"vgg16_tl_{TRAINING_RUN_ID}_{split}.h5")),
        tf.keras.callbacks.ProgbarLogger(
            count_mode = 'steps',
            stateful_metrics = None
        ),
        tf.keras.callbacks.EarlyStopping(
            monitor='val_accuracy',
            min_delta=0,
            patience=40,
            verbose=1,
            mode='auto',
            restore_best_weights=True
        )
    ]

    fit_params = {
                    "x": training_split_data,
                    "epochs": 105,
                    "callbacks": vgg16_tl_callbacks,
                    "validation_data": val_split_data,
                    "steps_per_epoch": 64,
                    "validation_steps": 10,
                    }
    
    preproc_func = tf.keras.applications.vgg16.preprocess_input
        
    with tf.device(TRAINING_DEVICE_NAME):

        vgg16_tl_training_history = train_model(vgg16_tl_model,
                                                fit_params=fit_params,
                                                preproc_func=preproc_func)
        
        _, pred_report = evaluate_model(vgg16_tl_model,
                                        vgg16_tl_training_history,
                                        fit_params=fit_params,
                                        preproc_func=preproc_func)
    
        vgg16_tl_cv_val_pred.append(pred_report)

    split += 1

calculate_cv_mean_metrics(vgg16_tl_cv_val_pred)

***************************
* Started at 12.171875... *
***************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 17.7204 - accuracy: 0.1318
Epoch 1: val_accuracy improved from -inf to 0.14375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 47s 664ms/step - loss: 17.7204 - accuracy: 0.1318 - val_loss: 2.7921 - val_accuracy: 0.1437
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.1072 - accuracy: 0.2393
Epoch 2: val_accuracy improved from 0.14375 to 0.29375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 42s 653ms/step - loss: 2.1072 - accuracy: 0.2393 - val_loss: 1.9249 - val_accuracy: 0.2937
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.0038 - accuracy: 0.2754
Epoch 3: val_accuracy did not improve from 0.29375
64/64 [==============================] - 19s 290ms/step - loss: 2.0038 - accuracy: 0.2754 - val_loss: 2.0032 - val_accuracy: 0.2500
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 1.9135 - accuracy: 0.2939
Epoch 4: val_accuracy did not improve from 0.29375
64/64 [==============================] - 13s 199ms/step - loss: 1.9135 - accuracy: 0.2939 - val_loss: 2.0673 - val_accuracy: 0.2250
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 1.9007 - accuracy: 0.2832
Epoch 5: val_accuracy improved from 0.29375 to 0.33750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 37s 569ms/step - loss: 1.9007 - accuracy: 0.2832 - val_loss: 1.7949 - val_accuracy: 0.3375
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 1.8988 - accuracy: 0.2920
Epoch 6: val_accuracy improved from 0.33750 to 0.38750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 39s 604ms/step - loss: 1.8988 - accuracy: 0.2920 - val_loss: 1.6587 - val_accuracy: 0.3875
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 1.8206 - accuracy: 0.3193
Epoch 7: val_accuracy did not improve from 0.38750
64/64 [==============================] - 13s 199ms/step - loss: 1.8206 - accuracy: 0.3193 - val_loss: 1.8346 - val_accuracy: 0.3313
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 1.8560 - accuracy: 0.3145
Epoch 8: val_accuracy did not improve from 0.38750
64/64 [==============================] - 13s 197ms/step - loss: 1.8560 - accuracy: 0.3145 - val_loss: 1.7644 - val_accuracy: 0.3063
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 1.8007 - accuracy: 0.3359
Epoch 9: val_accuracy did not improve from 0.38750
64/64 [==============================] - 12s 180ms/step - loss: 1.8007 - accuracy: 0.3359 - val_loss: 1.6989 - val_accuracy: 0.3438
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 1.7822 - accuracy: 0.3359
Epoch 10: val_accuracy did not improve from 0.38750
64/64 [==============================] - 12s 178ms/step - loss: 1.7822 - accuracy: 0.3359 - val_loss: 2.1629 - val_accuracy: 0.2500
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 1.8573 - accuracy: 0.3271
Epoch 11: val_accuracy did not improve from 0.38750
64/64 [==============================] - 11s 172ms/step - loss: 1.8573 - accuracy: 0.3271 - val_loss: 1.8467 - val_accuracy: 0.3438
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 1.7827 - accuracy: 0.3506
Epoch 12: val_accuracy did not improve from 0.38750
64/64 [==============================] - 10s 149ms/step - loss: 1.7827 - accuracy: 0.3506 - val_loss: 2.0629 - val_accuracy: 0.3063
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 1.8045 - accuracy: 0.3457
Epoch 13: val_accuracy improved from 0.38750 to 0.40625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 33s 518ms/step - loss: 1.8045 - accuracy: 0.3457 - val_loss: 1.5942 - val_accuracy: 0.4062
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 1.8070 - accuracy: 0.3525
Epoch 14: val_accuracy did not improve from 0.40625
64/64 [==============================] - 11s 170ms/step - loss: 1.8070 - accuracy: 0.3525 - val_loss: 1.6353 - val_accuracy: 0.3688
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 1.7093 - accuracy: 0.3467
Epoch 15: val_accuracy did not improve from 0.40625
64/64 [==============================] - 9s 141ms/step - loss: 1.7093 - accuracy: 0.3467 - val_loss: 1.7120 - val_accuracy: 0.3688
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 1.7342 - accuracy: 0.3555
Epoch 16: val_accuracy improved from 0.40625 to 0.41875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 32s 502ms/step - loss: 1.7342 - accuracy: 0.3555 - val_loss: 1.5410 - val_accuracy: 0.4187
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 1.6355 - accuracy: 0.3730
Epoch 17: val_accuracy did not improve from 0.41875
64/64 [==============================] - 11s 161ms/step - loss: 1.6355 - accuracy: 0.3730 - val_loss: 1.7243 - val_accuracy: 0.3438
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 1.7663 - accuracy: 0.3213
Epoch 18: val_accuracy did not improve from 0.41875
64/64 [==============================] - 9s 146ms/step - loss: 1.7663 - accuracy: 0.3213 - val_loss: 1.7547 - val_accuracy: 0.3500
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 1.7226 - accuracy: 0.3604
Epoch 19: val_accuracy did not improve from 0.41875
64/64 [==============================] - 9s 137ms/step - loss: 1.7226 - accuracy: 0.3604 - val_loss: 1.7196 - val_accuracy: 0.3750
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 1.6731 - accuracy: 0.3867
Epoch 20: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 128ms/step - loss: 1.6731 - accuracy: 0.3867 - val_loss: 1.6049 - val_accuracy: 0.4125
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 1.6336 - accuracy: 0.4150
Epoch 21: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 125ms/step - loss: 1.6336 - accuracy: 0.4150 - val_loss: 1.6809 - val_accuracy: 0.3500
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 1.6673 - accuracy: 0.3594
Epoch 22: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 126ms/step - loss: 1.6673 - accuracy: 0.3594 - val_loss: 1.9297 - val_accuracy: 0.3562
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 1.6387 - accuracy: 0.3936
Epoch 23: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 118ms/step - loss: 1.6387 - accuracy: 0.3936 - val_loss: 1.7772 - val_accuracy: 0.3812
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 1.6049 - accuracy: 0.4014
Epoch 24: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 118ms/step - loss: 1.6049 - accuracy: 0.4014 - val_loss: 1.6113 - val_accuracy: 0.3875
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 1.6542 - accuracy: 0.3633
Epoch 25: val_accuracy improved from 0.41875 to 0.43750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 30s 480ms/step - loss: 1.6542 - accuracy: 0.3633 - val_loss: 1.5523 - val_accuracy: 0.4375
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 1.6244 - accuracy: 0.3994
Epoch 26: val_accuracy did not improve from 0.43750
64/64 [==============================] - 8s 128ms/step - loss: 1.6244 - accuracy: 0.3994 - val_loss: 1.9127 - val_accuracy: 0.3562
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 1.6794 - accuracy: 0.4072
Epoch 27: val_accuracy improved from 0.43750 to 0.47500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 30s 475ms/step - loss: 1.6794 - accuracy: 0.4072 - val_loss: 1.5136 - val_accuracy: 0.4750
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 1.5675 - accuracy: 0.4424
Epoch 28: val_accuracy did not improve from 0.47500
64/64 [==============================] - 9s 129ms/step - loss: 1.5675 - accuracy: 0.4424 - val_loss: 1.5238 - val_accuracy: 0.4437
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.6885 - accuracy: 0.3779
Epoch 29: val_accuracy did not improve from 0.47500
64/64 [==============================] - 8s 120ms/step - loss: 1.6885 - accuracy: 0.3779 - val_loss: 1.5482 - val_accuracy: 0.3938
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.5719 - accuracy: 0.4287
Epoch 30: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 111ms/step - loss: 1.5719 - accuracy: 0.4287 - val_loss: 1.5013 - val_accuracy: 0.4500
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.5706 - accuracy: 0.4258
Epoch 31: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 112ms/step - loss: 1.5706 - accuracy: 0.4258 - val_loss: 1.6631 - val_accuracy: 0.4250
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.5842 - accuracy: 0.4395
Epoch 32: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 110ms/step - loss: 1.5842 - accuracy: 0.4395 - val_loss: 1.7439 - val_accuracy: 0.4250
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.5988 - accuracy: 0.4209
Epoch 33: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 109ms/step - loss: 1.5988 - accuracy: 0.4209 - val_loss: 1.7139 - val_accuracy: 0.4062
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.5942 - accuracy: 0.4385
Epoch 34: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 110ms/step - loss: 1.5942 - accuracy: 0.4385 - val_loss: 1.7520 - val_accuracy: 0.4062
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.6103 - accuracy: 0.4160
Epoch 35: val_accuracy improved from 0.47500 to 0.50625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 31s 490ms/step - loss: 1.6103 - accuracy: 0.4160 - val_loss: 1.4144 - val_accuracy: 0.5063
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.5277 - accuracy: 0.4600
Epoch 36: val_accuracy did not improve from 0.50625
64/64 [==============================] - 8s 112ms/step - loss: 1.5277 - accuracy: 0.4600 - val_loss: 1.7619 - val_accuracy: 0.4187
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.5708 - accuracy: 0.4316
Epoch 37: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 112ms/step - loss: 1.5708 - accuracy: 0.4316 - val_loss: 1.5884 - val_accuracy: 0.4437
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.6199 - accuracy: 0.4248
Epoch 38: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.6199 - accuracy: 0.4248 - val_loss: 1.6657 - val_accuracy: 0.4062
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.5616 - accuracy: 0.4404
Epoch 39: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.5616 - accuracy: 0.4404 - val_loss: 1.9636 - val_accuracy: 0.3250
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.5434 - accuracy: 0.4336
Epoch 40: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 108ms/step - loss: 1.5434 - accuracy: 0.4336 - val_loss: 1.6731 - val_accuracy: 0.4187
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.5703 - accuracy: 0.4326
Epoch 41: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 110ms/step - loss: 1.5703 - accuracy: 0.4326 - val_loss: 1.6319 - val_accuracy: 0.3938
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.5465 - accuracy: 0.4473
Epoch 42: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 108ms/step - loss: 1.5465 - accuracy: 0.4473 - val_loss: 1.5500 - val_accuracy: 0.4563
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.5903 - accuracy: 0.4404
Epoch 43: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.5903 - accuracy: 0.4404 - val_loss: 1.7327 - val_accuracy: 0.4062
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.5304 - accuracy: 0.4424
Epoch 44: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 105ms/step - loss: 1.5304 - accuracy: 0.4424 - val_loss: 1.5608 - val_accuracy: 0.4250
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.5687 - accuracy: 0.4404
Epoch 45: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 106ms/step - loss: 1.5687 - accuracy: 0.4404 - val_loss: 1.7591 - val_accuracy: 0.4000
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.5225 - accuracy: 0.4365
Epoch 46: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 106ms/step - loss: 1.5225 - accuracy: 0.4365 - val_loss: 1.6563 - val_accuracy: 0.4375
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.5803 - accuracy: 0.4248
Epoch 47: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 105ms/step - loss: 1.5803 - accuracy: 0.4248 - val_loss: 2.0241 - val_accuracy: 0.2937
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.5319 - accuracy: 0.4541
Epoch 48: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 104ms/step - loss: 1.5319 - accuracy: 0.4541 - val_loss: 1.6974 - val_accuracy: 0.4062
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.5196 - accuracy: 0.4512
Epoch 49: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 104ms/step - loss: 1.5196 - accuracy: 0.4512 - val_loss: 2.2677 - val_accuracy: 0.3250
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.5206 - accuracy: 0.4512
Epoch 50: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.5206 - accuracy: 0.4512 - val_loss: 1.8661 - val_accuracy: 0.4062
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.5237 - accuracy: 0.4629
Epoch 51: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 112ms/step - loss: 1.5237 - accuracy: 0.4629 - val_loss: 1.9110 - val_accuracy: 0.3250
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.4838 - accuracy: 0.4727
Epoch 52: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 111ms/step - loss: 1.4838 - accuracy: 0.4727 - val_loss: 1.6284 - val_accuracy: 0.4250
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.5545 - accuracy: 0.4199
Epoch 53: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 106ms/step - loss: 1.5545 - accuracy: 0.4199 - val_loss: 1.4564 - val_accuracy: 0.4750
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.4792 - accuracy: 0.4561
Epoch 54: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 108ms/step - loss: 1.4792 - accuracy: 0.4561 - val_loss: 1.5171 - val_accuracy: 0.4812
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.4632 - accuracy: 0.4512
Epoch 55: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.4632 - accuracy: 0.4512 - val_loss: 1.6384 - val_accuracy: 0.4250
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.4624 - accuracy: 0.4707
Epoch 56: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.4624 - accuracy: 0.4707 - val_loss: 1.7668 - val_accuracy: 0.4563
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.4761 - accuracy: 0.4561
Epoch 57: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 107ms/step - loss: 1.4761 - accuracy: 0.4561 - val_loss: 1.7704 - val_accuracy: 0.4187
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.4924 - accuracy: 0.4658
Epoch 58: val_accuracy improved from 0.50625 to 0.52500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 29s 460ms/step - loss: 1.4924 - accuracy: 0.4658 - val_loss: 1.5019 - val_accuracy: 0.5250
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.5566 - accuracy: 0.4346
Epoch 59: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.5566 - accuracy: 0.4346 - val_loss: 1.5978 - val_accuracy: 0.4625
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.5502 - accuracy: 0.4297
Epoch 60: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 106ms/step - loss: 1.5502 - accuracy: 0.4297 - val_loss: 2.0731 - val_accuracy: 0.3500
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.4650 - accuracy: 0.4648
Epoch 61: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4650 - accuracy: 0.4648 - val_loss: 1.4992 - val_accuracy: 0.5000
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.4647 - accuracy: 0.5039
Epoch 62: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 108ms/step - loss: 1.4647 - accuracy: 0.5039 - val_loss: 1.7177 - val_accuracy: 0.4500
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.5212 - accuracy: 0.4453
Epoch 63: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.5212 - accuracy: 0.4453 - val_loss: 1.5695 - val_accuracy: 0.5125
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.5157 - accuracy: 0.4736
Epoch 64: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.5157 - accuracy: 0.4736 - val_loss: 1.7630 - val_accuracy: 0.3938
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.5267 - accuracy: 0.4199
Epoch 65: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 106ms/step - loss: 1.5267 - accuracy: 0.4199 - val_loss: 1.5744 - val_accuracy: 0.4375
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.4379 - accuracy: 0.4883
Epoch 66: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 107ms/step - loss: 1.4379 - accuracy: 0.4883 - val_loss: 1.7750 - val_accuracy: 0.4437
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.4811 - accuracy: 0.4541
Epoch 67: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 106ms/step - loss: 1.4811 - accuracy: 0.4541 - val_loss: 1.6242 - val_accuracy: 0.4437
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.4020 - accuracy: 0.4971
Epoch 68: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4020 - accuracy: 0.4971 - val_loss: 1.7468 - val_accuracy: 0.4125
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.4997 - accuracy: 0.4697
Epoch 69: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4997 - accuracy: 0.4697 - val_loss: 1.9306 - val_accuracy: 0.4062
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.4636 - accuracy: 0.4795
Epoch 70: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4636 - accuracy: 0.4795 - val_loss: 1.5397 - val_accuracy: 0.4625
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.4838 - accuracy: 0.4678
Epoch 71: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4838 - accuracy: 0.4678 - val_loss: 1.4950 - val_accuracy: 0.4938
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.4245 - accuracy: 0.4893
Epoch 72: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4245 - accuracy: 0.4893 - val_loss: 1.5151 - val_accuracy: 0.4625
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.4605 - accuracy: 0.4707
Epoch 73: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4605 - accuracy: 0.4707 - val_loss: 1.5582 - val_accuracy: 0.4625
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.4142 - accuracy: 0.4883
Epoch 74: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4142 - accuracy: 0.4883 - val_loss: 1.7057 - val_accuracy: 0.4437
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.4426 - accuracy: 0.4717
Epoch 75: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4426 - accuracy: 0.4717 - val_loss: 1.6756 - val_accuracy: 0.4375
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.4706 - accuracy: 0.4619
Epoch 76: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4706 - accuracy: 0.4619 - val_loss: 1.6434 - val_accuracy: 0.4812
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.4382 - accuracy: 0.4863
Epoch 77: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 105ms/step - loss: 1.4382 - accuracy: 0.4863 - val_loss: 1.6838 - val_accuracy: 0.4563
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.3947 - accuracy: 0.5059
Epoch 78: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.3947 - accuracy: 0.5059 - val_loss: 1.8720 - val_accuracy: 0.4062
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.4470 - accuracy: 0.4678
Epoch 79: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4470 - accuracy: 0.4678 - val_loss: 1.4979 - val_accuracy: 0.5250
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.3930 - accuracy: 0.5156
Epoch 80: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 105ms/step - loss: 1.3930 - accuracy: 0.5156 - val_loss: 1.8121 - val_accuracy: 0.4125
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.4679 - accuracy: 0.4824
Epoch 81: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4679 - accuracy: 0.4824 - val_loss: 1.6343 - val_accuracy: 0.4500
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.4269 - accuracy: 0.4990
Epoch 82: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4269 - accuracy: 0.4990 - val_loss: 1.6011 - val_accuracy: 0.4313
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.4466 - accuracy: 0.5127
Epoch 83: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4466 - accuracy: 0.5127 - val_loss: 1.7736 - val_accuracy: 0.4563
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.4224 - accuracy: 0.4932
Epoch 84: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4224 - accuracy: 0.4932 - val_loss: 1.5448 - val_accuracy: 0.4625
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.4505 - accuracy: 0.4814
Epoch 85: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.4505 - accuracy: 0.4814 - val_loss: 1.6425 - val_accuracy: 0.4500
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.4151 - accuracy: 0.4941
Epoch 86: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4151 - accuracy: 0.4941 - val_loss: 1.3871 - val_accuracy: 0.5125
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.3500 - accuracy: 0.5020
Epoch 87: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 105ms/step - loss: 1.3500 - accuracy: 0.5020 - val_loss: 1.5208 - val_accuracy: 0.4563
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.4366 - accuracy: 0.4775
Epoch 88: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4366 - accuracy: 0.4775 - val_loss: 1.5407 - val_accuracy: 0.4812
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.4361 - accuracy: 0.4805
Epoch 89: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4361 - accuracy: 0.4805 - val_loss: 1.8043 - val_accuracy: 0.4062
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.4979 - accuracy: 0.4805
Epoch 90: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 105ms/step - loss: 1.4979 - accuracy: 0.4805 - val_loss: 1.4669 - val_accuracy: 0.4625
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.3780 - accuracy: 0.4980
Epoch 91: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.3780 - accuracy: 0.4980 - val_loss: 1.6739 - val_accuracy: 0.4688
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.3541 - accuracy: 0.4814
Epoch 92: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.3541 - accuracy: 0.4814 - val_loss: 1.5042 - val_accuracy: 0.4812
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.4046 - accuracy: 0.4961
Epoch 93: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 103ms/step - loss: 1.4046 - accuracy: 0.4961 - val_loss: 1.4637 - val_accuracy: 0.5125
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.4981 - accuracy: 0.4932
Epoch 94: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 102ms/step - loss: 1.4981 - accuracy: 0.4932 - val_loss: 1.4601 - val_accuracy: 0.4938
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.3733 - accuracy: 0.5381
Epoch 95: val_accuracy did not improve from 0.52500
64/64 [==============================] - 7s 104ms/step - loss: 1.3733 - accuracy: 0.5381 - val_loss: 1.4400 - val_accuracy: 0.5063
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.3952 - accuracy: 0.4912
Epoch 96: val_accuracy improved from 0.52500 to 0.55000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_1.h5
64/64 [==============================] - 31s 490ms/step - loss: 1.3952 - accuracy: 0.4912 - val_loss: 1.3826 - val_accuracy: 0.5500
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.3861 - accuracy: 0.5107
Epoch 97: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 102ms/step - loss: 1.3861 - accuracy: 0.5107 - val_loss: 1.6908 - val_accuracy: 0.4750
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.4213 - accuracy: 0.4863
Epoch 98: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 103ms/step - loss: 1.4213 - accuracy: 0.4863 - val_loss: 1.4225 - val_accuracy: 0.5063
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.5202 - accuracy: 0.4785
Epoch 99: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 104ms/step - loss: 1.5202 - accuracy: 0.4785 - val_loss: 1.5904 - val_accuracy: 0.4688
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.4948 - accuracy: 0.4775
Epoch 100: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 104ms/step - loss: 1.4948 - accuracy: 0.4775 - val_loss: 1.3256 - val_accuracy: 0.5000
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.4358 - accuracy: 0.4775
Epoch 101: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 104ms/step - loss: 1.4358 - accuracy: 0.4775 - val_loss: 1.3496 - val_accuracy: 0.4875
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.3475 - accuracy: 0.5166
Epoch 102: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 103ms/step - loss: 1.3475 - accuracy: 0.5166 - val_loss: 1.5244 - val_accuracy: 0.4437
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.4159 - accuracy: 0.4814
Epoch 103: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 103ms/step - loss: 1.4159 - accuracy: 0.4814 - val_loss: 1.5275 - val_accuracy: 0.4875
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4370 - accuracy: 0.4795
Epoch 104: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 103ms/step - loss: 1.4370 - accuracy: 0.4795 - val_loss: 1.5317 - val_accuracy: 0.4437
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.4831 - accuracy: 0.4590
Epoch 105: val_accuracy did not improve from 0.55000
64/64 [==============================] - 7s 103ms/step - loss: 1.4831 - accuracy: 0.4590 - val_loss: 1.4852 - val_accuracy: 0.4938


 ********* Training time: 1185.859375 s.


*****************
* Model Summary *
*****************
Model: "model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense (Dense)               (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 119,595,020
Non-trainable params: 14,714,688
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 20s 87ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.5966    0.9311    0.7273       305
anarrhichomenum     0.8261    0.5311    0.6465       322
   brevantherum     0.5946    0.1654    0.2588       266
      dulcamara     0.4483    0.1254    0.1960       311
   herposolanum     0.3778    0.4265    0.4007       279
     holophylla     0.4006    0.4858    0.4391       282
     lasiocarpa     0.9516    0.7540    0.8414       313
      melongena     0.7626    0.6242    0.6865       314
    micracantha     0.4686    0.3746    0.4164       299
         petota     0.2255    0.8600    0.3573       300
        solanum     0.2914    0.1678    0.2129       304
          torva     0.5833    0.0689    0.1232       305

       accuracy                         0.4633      3600
      macro avg     0.5439    0.4596    0.4422      3600
   weighted avg     0.5488    0.4633    0.4468      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*****************************
* Started at 1211.296875... *
*****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 16.3871 - accuracy: 0.1191
Epoch 1: val_accuracy improved from -inf to 0.18125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 28s 438ms/step - loss: 16.3871 - accuracy: 0.1191 - val_loss: 2.2480 - val_accuracy: 0.1813
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.1493 - accuracy: 0.2236
Epoch 2: val_accuracy improved from 0.18125 to 0.25000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 27s 418ms/step - loss: 2.1493 - accuracy: 0.2236 - val_loss: 2.0837 - val_accuracy: 0.2500
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.0542 - accuracy: 0.2520
Epoch 3: val_accuracy did not improve from 0.25000
64/64 [==============================] - 7s 100ms/step - loss: 2.0542 - accuracy: 0.2520 - val_loss: 2.2026 - val_accuracy: 0.2125
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 1.9634 - accuracy: 0.2900
Epoch 4: val_accuracy improved from 0.25000 to 0.26875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 27s 431ms/step - loss: 1.9634 - accuracy: 0.2900 - val_loss: 1.9547 - val_accuracy: 0.2688
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 1.8545 - accuracy: 0.3164
Epoch 5: val_accuracy did not improve from 0.26875
64/64 [==============================] - 7s 100ms/step - loss: 1.8545 - accuracy: 0.3164 - val_loss: 2.3121 - val_accuracy: 0.2500
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 1.9023 - accuracy: 0.3193
Epoch 6: val_accuracy did not improve from 0.26875
64/64 [==============================] - 7s 103ms/step - loss: 1.9023 - accuracy: 0.3193 - val_loss: 2.0366 - val_accuracy: 0.2562
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 1.8571 - accuracy: 0.3213
Epoch 7: val_accuracy did not improve from 0.26875
64/64 [==============================] - 7s 103ms/step - loss: 1.8571 - accuracy: 0.3213 - val_loss: 1.9992 - val_accuracy: 0.2625
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 1.8012 - accuracy: 0.3320
Epoch 8: val_accuracy improved from 0.26875 to 0.30000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 27s 433ms/step - loss: 1.8012 - accuracy: 0.3320 - val_loss: 1.8987 - val_accuracy: 0.3000
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 1.7224 - accuracy: 0.3896
Epoch 9: val_accuracy improved from 0.30000 to 0.32500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 28s 437ms/step - loss: 1.7224 - accuracy: 0.3896 - val_loss: 1.8684 - val_accuracy: 0.3250
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 1.7510 - accuracy: 0.3613
Epoch 10: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 103ms/step - loss: 1.7510 - accuracy: 0.3613 - val_loss: 2.0537 - val_accuracy: 0.2937
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 1.7817 - accuracy: 0.3623
Epoch 11: val_accuracy improved from 0.32500 to 0.36875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 29s 459ms/step - loss: 1.7817 - accuracy: 0.3623 - val_loss: 1.9210 - val_accuracy: 0.3688
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 1.7672 - accuracy: 0.3535
Epoch 12: val_accuracy improved from 0.36875 to 0.40625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 28s 432ms/step - loss: 1.7672 - accuracy: 0.3535 - val_loss: 1.7583 - val_accuracy: 0.4062
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 1.6562 - accuracy: 0.3906
Epoch 13: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6562 - accuracy: 0.3906 - val_loss: 1.7859 - val_accuracy: 0.3750
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 1.7044 - accuracy: 0.3770
Epoch 14: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 103ms/step - loss: 1.7044 - accuracy: 0.3770 - val_loss: 2.0178 - val_accuracy: 0.3250
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 1.7101 - accuracy: 0.3691
Epoch 15: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.7101 - accuracy: 0.3691 - val_loss: 1.7343 - val_accuracy: 0.3187
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 1.7359 - accuracy: 0.3789
Epoch 16: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 104ms/step - loss: 1.7359 - accuracy: 0.3789 - val_loss: 2.1638 - val_accuracy: 0.3000
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 1.7549 - accuracy: 0.3711
Epoch 17: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 103ms/step - loss: 1.7549 - accuracy: 0.3711 - val_loss: 1.8435 - val_accuracy: 0.2937
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 1.6975 - accuracy: 0.3770
Epoch 18: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6975 - accuracy: 0.3770 - val_loss: 1.8149 - val_accuracy: 0.3375
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 1.6036 - accuracy: 0.4307
Epoch 19: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6036 - accuracy: 0.4307 - val_loss: 1.8716 - val_accuracy: 0.3688
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 1.6699 - accuracy: 0.4023
Epoch 20: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6699 - accuracy: 0.4023 - val_loss: 2.0249 - val_accuracy: 0.3438
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 1.6352 - accuracy: 0.4033
Epoch 21: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6352 - accuracy: 0.4033 - val_loss: 1.7445 - val_accuracy: 0.3750
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 1.6469 - accuracy: 0.4102
Epoch 22: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 101ms/step - loss: 1.6469 - accuracy: 0.4102 - val_loss: 1.8729 - val_accuracy: 0.3500
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 1.6306 - accuracy: 0.4023
Epoch 23: val_accuracy did not improve from 0.40625
64/64 [==============================] - 6s 101ms/step - loss: 1.6306 - accuracy: 0.4023 - val_loss: 1.6248 - val_accuracy: 0.4062
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 1.5786 - accuracy: 0.4189
Epoch 24: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.5786 - accuracy: 0.4189 - val_loss: 1.8130 - val_accuracy: 0.4062
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 1.5609 - accuracy: 0.4453
Epoch 25: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.5609 - accuracy: 0.4453 - val_loss: 2.0323 - val_accuracy: 0.3187
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 1.6195 - accuracy: 0.4131
Epoch 26: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 102ms/step - loss: 1.6195 - accuracy: 0.4131 - val_loss: 1.5706 - val_accuracy: 0.3875
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 1.5749 - accuracy: 0.4336
Epoch 27: val_accuracy did not improve from 0.40625
64/64 [==============================] - 6s 101ms/step - loss: 1.5749 - accuracy: 0.4336 - val_loss: 1.9523 - val_accuracy: 0.3625
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 1.6169 - accuracy: 0.4355
Epoch 28: val_accuracy improved from 0.40625 to 0.43750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 28s 444ms/step - loss: 1.6169 - accuracy: 0.4355 - val_loss: 1.6352 - val_accuracy: 0.4375
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.5686 - accuracy: 0.4277
Epoch 29: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 100ms/step - loss: 1.5686 - accuracy: 0.4277 - val_loss: 1.6465 - val_accuracy: 0.4313
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.4912 - accuracy: 0.4766
Epoch 30: val_accuracy improved from 0.43750 to 0.46250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 29s 454ms/step - loss: 1.4912 - accuracy: 0.4766 - val_loss: 1.4947 - val_accuracy: 0.4625
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.5191 - accuracy: 0.4492
Epoch 31: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 101ms/step - loss: 1.5191 - accuracy: 0.4492 - val_loss: 1.7154 - val_accuracy: 0.4187
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.5092 - accuracy: 0.4639
Epoch 32: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 103ms/step - loss: 1.5092 - accuracy: 0.4639 - val_loss: 1.5895 - val_accuracy: 0.4313
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.5359 - accuracy: 0.4492
Epoch 33: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 102ms/step - loss: 1.5359 - accuracy: 0.4492 - val_loss: 1.5994 - val_accuracy: 0.4062
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.5372 - accuracy: 0.4551
Epoch 34: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 103ms/step - loss: 1.5372 - accuracy: 0.4551 - val_loss: 1.8490 - val_accuracy: 0.3438
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.5305 - accuracy: 0.4531
Epoch 35: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 102ms/step - loss: 1.5305 - accuracy: 0.4531 - val_loss: 1.5525 - val_accuracy: 0.4375
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.5019 - accuracy: 0.4541
Epoch 36: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 103ms/step - loss: 1.5019 - accuracy: 0.4541 - val_loss: 1.9044 - val_accuracy: 0.4187
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.5379 - accuracy: 0.4600
Epoch 37: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 102ms/step - loss: 1.5379 - accuracy: 0.4600 - val_loss: 1.5478 - val_accuracy: 0.4500
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.5357 - accuracy: 0.4619
Epoch 38: val_accuracy improved from 0.46250 to 0.50000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 29s 456ms/step - loss: 1.5357 - accuracy: 0.4619 - val_loss: 1.5461 - val_accuracy: 0.5000
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.4690 - accuracy: 0.4600
Epoch 39: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 101ms/step - loss: 1.4690 - accuracy: 0.4600 - val_loss: 1.6265 - val_accuracy: 0.4125
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.4988 - accuracy: 0.4805
Epoch 40: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4988 - accuracy: 0.4805 - val_loss: 1.5442 - val_accuracy: 0.4812
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.4779 - accuracy: 0.4785
Epoch 41: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4779 - accuracy: 0.4785 - val_loss: 1.6920 - val_accuracy: 0.4125
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.4609 - accuracy: 0.4912
Epoch 42: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4609 - accuracy: 0.4912 - val_loss: 1.5677 - val_accuracy: 0.4375
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.5125 - accuracy: 0.4639
Epoch 43: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.5125 - accuracy: 0.4639 - val_loss: 1.5863 - val_accuracy: 0.4250
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.6024 - accuracy: 0.4238
Epoch 44: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.6024 - accuracy: 0.4238 - val_loss: 1.8969 - val_accuracy: 0.3688
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.4594 - accuracy: 0.4795
Epoch 45: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4594 - accuracy: 0.4795 - val_loss: 1.5980 - val_accuracy: 0.4187
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.4983 - accuracy: 0.4531
Epoch 46: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4983 - accuracy: 0.4531 - val_loss: 1.6318 - val_accuracy: 0.4313
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.5174 - accuracy: 0.4746
Epoch 47: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.5174 - accuracy: 0.4746 - val_loss: 1.6752 - val_accuracy: 0.4313
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.4342 - accuracy: 0.5156
Epoch 48: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4342 - accuracy: 0.5156 - val_loss: 1.4710 - val_accuracy: 0.4938
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.4328 - accuracy: 0.4795
Epoch 49: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4328 - accuracy: 0.4795 - val_loss: 1.5878 - val_accuracy: 0.4563
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.4426 - accuracy: 0.5039
Epoch 50: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4426 - accuracy: 0.5039 - val_loss: 1.8941 - val_accuracy: 0.4000
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.4973 - accuracy: 0.4834
Epoch 51: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4973 - accuracy: 0.4834 - val_loss: 1.4547 - val_accuracy: 0.4688
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.3864 - accuracy: 0.5225
Epoch 52: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.3864 - accuracy: 0.5225 - val_loss: 1.6489 - val_accuracy: 0.4375
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.4612 - accuracy: 0.4951
Epoch 53: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4612 - accuracy: 0.4951 - val_loss: 1.5791 - val_accuracy: 0.4688
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.4513 - accuracy: 0.4893
Epoch 54: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4513 - accuracy: 0.4893 - val_loss: 1.5249 - val_accuracy: 0.5000
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.4383 - accuracy: 0.4902
Epoch 55: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4383 - accuracy: 0.4902 - val_loss: 1.4272 - val_accuracy: 0.4875
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.4596 - accuracy: 0.4844
Epoch 56: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4596 - accuracy: 0.4844 - val_loss: 1.5880 - val_accuracy: 0.4375
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.3679 - accuracy: 0.5186
Epoch 57: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.3679 - accuracy: 0.5186 - val_loss: 1.4927 - val_accuracy: 0.4750
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.4127 - accuracy: 0.5049
Epoch 58: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4127 - accuracy: 0.5049 - val_loss: 1.4844 - val_accuracy: 0.4812
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.4553 - accuracy: 0.4756
Epoch 59: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.4553 - accuracy: 0.4756 - val_loss: 1.6643 - val_accuracy: 0.4062
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.3804 - accuracy: 0.5146
Epoch 60: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.3804 - accuracy: 0.5146 - val_loss: 1.6462 - val_accuracy: 0.4250
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.3940 - accuracy: 0.5176
Epoch 61: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.3940 - accuracy: 0.5176 - val_loss: 1.5588 - val_accuracy: 0.4688
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.4338 - accuracy: 0.5029
Epoch 62: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 103ms/step - loss: 1.4338 - accuracy: 0.5029 - val_loss: 1.4736 - val_accuracy: 0.4625
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.3177 - accuracy: 0.5371
Epoch 63: val_accuracy did not improve from 0.50000
64/64 [==============================] - 7s 102ms/step - loss: 1.3177 - accuracy: 0.5371 - val_loss: 1.6220 - val_accuracy: 0.4625
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.3748 - accuracy: 0.4990
Epoch 64: val_accuracy improved from 0.50000 to 0.50625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 31s 482ms/step - loss: 1.3748 - accuracy: 0.4990 - val_loss: 1.4365 - val_accuracy: 0.5063
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.3946 - accuracy: 0.5059
Epoch 65: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 101ms/step - loss: 1.3946 - accuracy: 0.5059 - val_loss: 1.5580 - val_accuracy: 0.4500
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.4763 - accuracy: 0.4746
Epoch 66: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 104ms/step - loss: 1.4763 - accuracy: 0.4746 - val_loss: 1.4200 - val_accuracy: 0.4563
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.4183 - accuracy: 0.4912
Epoch 67: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4183 - accuracy: 0.4912 - val_loss: 1.5226 - val_accuracy: 0.4625
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.4139 - accuracy: 0.4980
Epoch 68: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4139 - accuracy: 0.4980 - val_loss: 2.0720 - val_accuracy: 0.3625
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.4706 - accuracy: 0.4883
Epoch 69: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4706 - accuracy: 0.4883 - val_loss: 1.5608 - val_accuracy: 0.4812
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.4756 - accuracy: 0.4941
Epoch 70: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4756 - accuracy: 0.4941 - val_loss: 1.4504 - val_accuracy: 0.4688
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.4369 - accuracy: 0.4971
Epoch 71: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4369 - accuracy: 0.4971 - val_loss: 1.4509 - val_accuracy: 0.4812
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.3476 - accuracy: 0.5381
Epoch 72: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.3476 - accuracy: 0.5381 - val_loss: 1.6290 - val_accuracy: 0.4812
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.3081 - accuracy: 0.5459
Epoch 73: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.3081 - accuracy: 0.5459 - val_loss: 1.5180 - val_accuracy: 0.5000
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.4072 - accuracy: 0.5010
Epoch 74: val_accuracy improved from 0.50625 to 0.51250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 30s 472ms/step - loss: 1.4072 - accuracy: 0.5010 - val_loss: 1.4425 - val_accuracy: 0.5125
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.3383 - accuracy: 0.5225
Epoch 75: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.3383 - accuracy: 0.5225 - val_loss: 1.5875 - val_accuracy: 0.4500
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.3516 - accuracy: 0.5293
Epoch 76: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 103ms/step - loss: 1.3516 - accuracy: 0.5293 - val_loss: 1.4397 - val_accuracy: 0.5063
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.3915 - accuracy: 0.5244
Epoch 77: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.3915 - accuracy: 0.5244 - val_loss: 1.6090 - val_accuracy: 0.4375
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.2901 - accuracy: 0.5430
Epoch 78: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.2901 - accuracy: 0.5430 - val_loss: 1.4321 - val_accuracy: 0.4500
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.3121 - accuracy: 0.5322
Epoch 79: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 103ms/step - loss: 1.3121 - accuracy: 0.5322 - val_loss: 1.5119 - val_accuracy: 0.4875
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.3120 - accuracy: 0.5352
Epoch 80: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 103ms/step - loss: 1.3120 - accuracy: 0.5352 - val_loss: 1.6427 - val_accuracy: 0.4625
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.3788 - accuracy: 0.5068
Epoch 81: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.3788 - accuracy: 0.5068 - val_loss: 1.5633 - val_accuracy: 0.4500
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.3722 - accuracy: 0.5234
Epoch 82: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.3722 - accuracy: 0.5234 - val_loss: 1.3944 - val_accuracy: 0.5125
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.3339 - accuracy: 0.5312
Epoch 83: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 102ms/step - loss: 1.3339 - accuracy: 0.5312 - val_loss: 1.5706 - val_accuracy: 0.4375
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.3736 - accuracy: 0.5225
Epoch 84: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 103ms/step - loss: 1.3736 - accuracy: 0.5225 - val_loss: 1.6020 - val_accuracy: 0.4375
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.4229 - accuracy: 0.5127
Epoch 85: val_accuracy improved from 0.51250 to 0.54375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_2.h5
64/64 [==============================] - 29s 456ms/step - loss: 1.4229 - accuracy: 0.5127 - val_loss: 1.2757 - val_accuracy: 0.5437
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.3488 - accuracy: 0.5000
Epoch 86: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.3488 - accuracy: 0.5000 - val_loss: 1.3399 - val_accuracy: 0.5125
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.3441 - accuracy: 0.5361
Epoch 87: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3441 - accuracy: 0.5361 - val_loss: 1.4370 - val_accuracy: 0.4875
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.3096 - accuracy: 0.5381
Epoch 88: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.3096 - accuracy: 0.5381 - val_loss: 1.4781 - val_accuracy: 0.4688
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.3485 - accuracy: 0.5410
Epoch 89: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3485 - accuracy: 0.5410 - val_loss: 1.4840 - val_accuracy: 0.4812
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.3246 - accuracy: 0.5205
Epoch 90: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3246 - accuracy: 0.5205 - val_loss: 1.3527 - val_accuracy: 0.4500
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.3013 - accuracy: 0.5498
Epoch 91: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3013 - accuracy: 0.5498 - val_loss: 1.4252 - val_accuracy: 0.5125
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.2956 - accuracy: 0.5283
Epoch 92: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.2956 - accuracy: 0.5283 - val_loss: 1.5071 - val_accuracy: 0.5000
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.3344 - accuracy: 0.5234
Epoch 93: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3344 - accuracy: 0.5234 - val_loss: 1.4249 - val_accuracy: 0.4938
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.3531 - accuracy: 0.5205
Epoch 94: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3531 - accuracy: 0.5205 - val_loss: 1.6195 - val_accuracy: 0.4563
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.3555 - accuracy: 0.5283
Epoch 95: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3555 - accuracy: 0.5283 - val_loss: 1.3269 - val_accuracy: 0.5250
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.3656 - accuracy: 0.5381
Epoch 96: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3656 - accuracy: 0.5381 - val_loss: 1.8258 - val_accuracy: 0.4375
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.3899 - accuracy: 0.5176
Epoch 97: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3899 - accuracy: 0.5176 - val_loss: 1.4017 - val_accuracy: 0.4625
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.3450 - accuracy: 0.5371
Epoch 98: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3450 - accuracy: 0.5371 - val_loss: 1.2905 - val_accuracy: 0.5375
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.2814 - accuracy: 0.5479
Epoch 99: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.2814 - accuracy: 0.5479 - val_loss: 1.7082 - val_accuracy: 0.4625
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.2784 - accuracy: 0.5479
Epoch 100: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.2784 - accuracy: 0.5479 - val_loss: 1.4784 - val_accuracy: 0.4875
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.2793 - accuracy: 0.5596
Epoch 101: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.2793 - accuracy: 0.5596 - val_loss: 1.2870 - val_accuracy: 0.5250
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.3453 - accuracy: 0.5410
Epoch 102: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3453 - accuracy: 0.5410 - val_loss: 1.3056 - val_accuracy: 0.5250
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.3055 - accuracy: 0.5410
Epoch 103: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3055 - accuracy: 0.5410 - val_loss: 1.6197 - val_accuracy: 0.4750
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.2452 - accuracy: 0.5439
Epoch 104: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.2452 - accuracy: 0.5439 - val_loss: 1.3508 - val_accuracy: 0.5250
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.3001 - accuracy: 0.5625
Epoch 105: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.3001 - accuracy: 0.5625 - val_loss: 1.5886 - val_accuracy: 0.4250


 ********* Training time: 1149.75 s.


*****************
* Model Summary *
*****************
Model: "model_1"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense_1 (Dense)             (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 119,595,020
Non-trainable params: 14,714,688
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 5s 24ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.9266    0.5921    0.7225       277
anarrhichomenum     0.4472    0.9270    0.6033       315
   brevantherum     0.4957    0.3779    0.4288       307
      dulcamara     0.3161    0.2054    0.2490       297
   herposolanum     0.3324    0.8056    0.4706       288
     holophylla     0.5339    0.4685    0.4991       286
     lasiocarpa     0.7391    0.8662    0.7977       314
      melongena     0.8333    0.1546    0.2609       291
    micracantha     0.5922    0.3630    0.4501       292
         petota     0.4645    0.5579    0.5069       328
        solanum     0.3371    0.3947    0.3636       304
          torva     0.4651    0.0664    0.1163       301

       accuracy                         0.4847      3600
      macro avg     0.5403    0.4816    0.4557      3600
   weighted avg     0.5373    0.4847    0.4566      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*****************************
* Started at 2373.234375... *
*****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 13.9492 - accuracy: 0.1250
Epoch 1: val_accuracy improved from -inf to 0.17500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 26s 410ms/step - loss: 13.9492 - accuracy: 0.1250 - val_loss: 2.2366 - val_accuracy: 0.1750
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.2193 - accuracy: 0.2012
Epoch 2: val_accuracy improved from 0.17500 to 0.29375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 24s 370ms/step - loss: 2.2193 - accuracy: 0.2012 - val_loss: 1.9254 - val_accuracy: 0.2937
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.0137 - accuracy: 0.2715
Epoch 3: val_accuracy improved from 0.29375 to 0.33125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 27s 419ms/step - loss: 2.0137 - accuracy: 0.2715 - val_loss: 1.7977 - val_accuracy: 0.3313
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 1.9923 - accuracy: 0.2549
Epoch 4: val_accuracy improved from 0.33125 to 0.36875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 26s 402ms/step - loss: 1.9923 - accuracy: 0.2549 - val_loss: 1.6853 - val_accuracy: 0.3688
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 1.9293 - accuracy: 0.3037
Epoch 5: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 101ms/step - loss: 1.9293 - accuracy: 0.3037 - val_loss: 1.7096 - val_accuracy: 0.3438
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 1.8189 - accuracy: 0.3174
Epoch 6: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 103ms/step - loss: 1.8189 - accuracy: 0.3174 - val_loss: 1.6366 - val_accuracy: 0.3562
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 1.9481 - accuracy: 0.2930
Epoch 7: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 102ms/step - loss: 1.9481 - accuracy: 0.2930 - val_loss: 1.7131 - val_accuracy: 0.3688
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 1.9226 - accuracy: 0.2998
Epoch 8: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 103ms/step - loss: 1.9226 - accuracy: 0.2998 - val_loss: 1.8304 - val_accuracy: 0.2500
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 1.8140 - accuracy: 0.3350
Epoch 9: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 102ms/step - loss: 1.8140 - accuracy: 0.3350 - val_loss: 1.7349 - val_accuracy: 0.3625
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 1.8789 - accuracy: 0.3223
Epoch 10: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 103ms/step - loss: 1.8789 - accuracy: 0.3223 - val_loss: 1.6202 - val_accuracy: 0.3562
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 1.7481 - accuracy: 0.3447
Epoch 11: val_accuracy did not improve from 0.36875
64/64 [==============================] - 7s 102ms/step - loss: 1.7481 - accuracy: 0.3447 - val_loss: 1.8148 - val_accuracy: 0.2688
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 1.8045 - accuracy: 0.3398
Epoch 12: val_accuracy improved from 0.36875 to 0.41250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 27s 420ms/step - loss: 1.8045 - accuracy: 0.3398 - val_loss: 1.5289 - val_accuracy: 0.4125
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 1.7600 - accuracy: 0.3623
Epoch 13: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 100ms/step - loss: 1.7600 - accuracy: 0.3623 - val_loss: 1.6841 - val_accuracy: 0.3812
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 1.7457 - accuracy: 0.3633
Epoch 14: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 101ms/step - loss: 1.7457 - accuracy: 0.3633 - val_loss: 1.5344 - val_accuracy: 0.3812
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 1.7504 - accuracy: 0.3721
Epoch 15: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 101ms/step - loss: 1.7504 - accuracy: 0.3721 - val_loss: 1.6686 - val_accuracy: 0.3750
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 1.7211 - accuracy: 0.3535
Epoch 16: val_accuracy did not improve from 0.41250
64/64 [==============================] - 6s 100ms/step - loss: 1.7211 - accuracy: 0.3535 - val_loss: 1.6678 - val_accuracy: 0.3500
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 1.6691 - accuracy: 0.3662
Epoch 17: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 102ms/step - loss: 1.6691 - accuracy: 0.3662 - val_loss: 1.7364 - val_accuracy: 0.3438
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 1.6798 - accuracy: 0.3916
Epoch 18: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 102ms/step - loss: 1.6798 - accuracy: 0.3916 - val_loss: 1.5583 - val_accuracy: 0.4062
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 1.7385 - accuracy: 0.3672
Epoch 19: val_accuracy improved from 0.41250 to 0.45625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 24s 371ms/step - loss: 1.7385 - accuracy: 0.3672 - val_loss: 1.4914 - val_accuracy: 0.4563
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 1.6910 - accuracy: 0.4043
Epoch 20: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 101ms/step - loss: 1.6910 - accuracy: 0.4043 - val_loss: 1.4718 - val_accuracy: 0.4313
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 1.7485 - accuracy: 0.3525
Epoch 21: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 103ms/step - loss: 1.7485 - accuracy: 0.3525 - val_loss: 1.5568 - val_accuracy: 0.3438
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 1.6549 - accuracy: 0.3770
Epoch 22: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 103ms/step - loss: 1.6549 - accuracy: 0.3770 - val_loss: 1.6153 - val_accuracy: 0.3438
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 1.6582 - accuracy: 0.3955
Epoch 23: val_accuracy improved from 0.45625 to 0.46250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 27s 419ms/step - loss: 1.6582 - accuracy: 0.3955 - val_loss: 1.3921 - val_accuracy: 0.4625
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 1.6832 - accuracy: 0.3965
Epoch 24: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 100ms/step - loss: 1.6832 - accuracy: 0.3965 - val_loss: 1.5592 - val_accuracy: 0.4000
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 1.6748 - accuracy: 0.3750
Epoch 25: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 102ms/step - loss: 1.6748 - accuracy: 0.3750 - val_loss: 1.5604 - val_accuracy: 0.4062
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 1.6247 - accuracy: 0.4102
Epoch 26: val_accuracy did not improve from 0.46250
64/64 [==============================] - 7s 102ms/step - loss: 1.6247 - accuracy: 0.4102 - val_loss: 1.8119 - val_accuracy: 0.2625
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 1.6456 - accuracy: 0.3799
Epoch 27: val_accuracy improved from 0.46250 to 0.48125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 26s 403ms/step - loss: 1.6456 - accuracy: 0.3799 - val_loss: 1.4013 - val_accuracy: 0.4812
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 1.6754 - accuracy: 0.3682
Epoch 28: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 101ms/step - loss: 1.6754 - accuracy: 0.3682 - val_loss: 1.4540 - val_accuracy: 0.4625
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.7488 - accuracy: 0.3691
Epoch 29: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 103ms/step - loss: 1.7488 - accuracy: 0.3691 - val_loss: 1.7092 - val_accuracy: 0.3375
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.6817 - accuracy: 0.3770
Epoch 30: val_accuracy improved from 0.48125 to 0.51875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 29s 463ms/step - loss: 1.6817 - accuracy: 0.3770 - val_loss: 1.3814 - val_accuracy: 0.5188
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.6031 - accuracy: 0.4004
Epoch 31: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 101ms/step - loss: 1.6031 - accuracy: 0.4004 - val_loss: 1.5201 - val_accuracy: 0.4625
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.6238 - accuracy: 0.4189
Epoch 32: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.6238 - accuracy: 0.4189 - val_loss: 1.4361 - val_accuracy: 0.4688
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.5343 - accuracy: 0.4248
Epoch 33: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.5343 - accuracy: 0.4248 - val_loss: 1.4479 - val_accuracy: 0.4750
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.6213 - accuracy: 0.4004
Epoch 34: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.6213 - accuracy: 0.4004 - val_loss: 1.3622 - val_accuracy: 0.5063
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.5877 - accuracy: 0.4248
Epoch 35: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 103ms/step - loss: 1.5877 - accuracy: 0.4248 - val_loss: 1.5495 - val_accuracy: 0.4688
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.6159 - accuracy: 0.4199
Epoch 36: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.6159 - accuracy: 0.4199 - val_loss: 1.6466 - val_accuracy: 0.4688
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.6216 - accuracy: 0.3916
Epoch 37: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.6216 - accuracy: 0.3916 - val_loss: 1.3399 - val_accuracy: 0.5000
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.5433 - accuracy: 0.4473
Epoch 38: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 103ms/step - loss: 1.5433 - accuracy: 0.4473 - val_loss: 1.3587 - val_accuracy: 0.4938
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.6331 - accuracy: 0.4014
Epoch 39: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.6331 - accuracy: 0.4014 - val_loss: 1.3785 - val_accuracy: 0.4625
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.5819 - accuracy: 0.4189
Epoch 40: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 102ms/step - loss: 1.5819 - accuracy: 0.4189 - val_loss: 1.3693 - val_accuracy: 0.5188
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.5481 - accuracy: 0.4258
Epoch 41: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 101ms/step - loss: 1.5481 - accuracy: 0.4258 - val_loss: 1.5886 - val_accuracy: 0.3938
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.6224 - accuracy: 0.4248
Epoch 42: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 103ms/step - loss: 1.6224 - accuracy: 0.4248 - val_loss: 1.4134 - val_accuracy: 0.5125
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.6149 - accuracy: 0.4111
Epoch 43: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 103ms/step - loss: 1.6149 - accuracy: 0.4111 - val_loss: 1.3886 - val_accuracy: 0.4563
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.5371 - accuracy: 0.4365
Epoch 44: val_accuracy did not improve from 0.51875
64/64 [==============================] - 7s 103ms/step - loss: 1.5371 - accuracy: 0.4365 - val_loss: 1.3166 - val_accuracy: 0.5000
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.5782 - accuracy: 0.4326
Epoch 45: val_accuracy improved from 0.51875 to 0.53125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 28s 447ms/step - loss: 1.5782 - accuracy: 0.4326 - val_loss: 1.3194 - val_accuracy: 0.5312
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.5697 - accuracy: 0.4473
Epoch 46: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 101ms/step - loss: 1.5697 - accuracy: 0.4473 - val_loss: 1.3254 - val_accuracy: 0.4875
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.5565 - accuracy: 0.4336
Epoch 47: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 104ms/step - loss: 1.5565 - accuracy: 0.4336 - val_loss: 1.3525 - val_accuracy: 0.5063
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.5202 - accuracy: 0.4365
Epoch 48: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5202 - accuracy: 0.4365 - val_loss: 1.6206 - val_accuracy: 0.3875
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.5129 - accuracy: 0.4619
Epoch 49: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5129 - accuracy: 0.4619 - val_loss: 1.3256 - val_accuracy: 0.4938
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.5595 - accuracy: 0.4307
Epoch 50: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5595 - accuracy: 0.4307 - val_loss: 1.3212 - val_accuracy: 0.4812
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.6031 - accuracy: 0.4297
Epoch 51: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.6031 - accuracy: 0.4297 - val_loss: 1.3493 - val_accuracy: 0.5063
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.5333 - accuracy: 0.4512
Epoch 52: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5333 - accuracy: 0.4512 - val_loss: 1.3619 - val_accuracy: 0.5125
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.5518 - accuracy: 0.4521
Epoch 53: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5518 - accuracy: 0.4521 - val_loss: 1.4650 - val_accuracy: 0.4750
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.5379 - accuracy: 0.4434
Epoch 54: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5379 - accuracy: 0.4434 - val_loss: 1.6789 - val_accuracy: 0.4250
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.5692 - accuracy: 0.4121
Epoch 55: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5692 - accuracy: 0.4121 - val_loss: 1.4892 - val_accuracy: 0.4375
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.4995 - accuracy: 0.4600
Epoch 56: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.4995 - accuracy: 0.4600 - val_loss: 1.3126 - val_accuracy: 0.5312
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.5580 - accuracy: 0.4443
Epoch 57: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5580 - accuracy: 0.4443 - val_loss: 1.7278 - val_accuracy: 0.3938
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.5308 - accuracy: 0.4375
Epoch 58: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5308 - accuracy: 0.4375 - val_loss: 1.3994 - val_accuracy: 0.4437
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.5601 - accuracy: 0.4277
Epoch 59: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5601 - accuracy: 0.4277 - val_loss: 1.4291 - val_accuracy: 0.4437
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.5991 - accuracy: 0.4248
Epoch 60: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5991 - accuracy: 0.4248 - val_loss: 1.4041 - val_accuracy: 0.4688
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.5132 - accuracy: 0.4385
Epoch 61: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5132 - accuracy: 0.4385 - val_loss: 1.3830 - val_accuracy: 0.4437
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.5690 - accuracy: 0.4346
Epoch 62: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5690 - accuracy: 0.4346 - val_loss: 1.3565 - val_accuracy: 0.4875
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.5529 - accuracy: 0.4336
Epoch 63: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5529 - accuracy: 0.4336 - val_loss: 1.2563 - val_accuracy: 0.4938
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.5526 - accuracy: 0.4395
Epoch 64: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5526 - accuracy: 0.4395 - val_loss: 1.4361 - val_accuracy: 0.4250
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.5329 - accuracy: 0.4336
Epoch 65: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5329 - accuracy: 0.4336 - val_loss: 1.4495 - val_accuracy: 0.4563
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.4661 - accuracy: 0.4766
Epoch 66: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.4661 - accuracy: 0.4766 - val_loss: 1.4157 - val_accuracy: 0.4625
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.5186 - accuracy: 0.4609
Epoch 67: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5186 - accuracy: 0.4609 - val_loss: 1.2997 - val_accuracy: 0.5188
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.5101 - accuracy: 0.4746
Epoch 68: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.5101 - accuracy: 0.4746 - val_loss: 1.3724 - val_accuracy: 0.4875
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.5192 - accuracy: 0.4561
Epoch 69: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5192 - accuracy: 0.4561 - val_loss: 1.5172 - val_accuracy: 0.4000
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.6428 - accuracy: 0.4229
Epoch 70: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 102ms/step - loss: 1.6428 - accuracy: 0.4229 - val_loss: 1.6652 - val_accuracy: 0.3688
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.5165 - accuracy: 0.4531
Epoch 71: val_accuracy did not improve from 0.53125
64/64 [==============================] - 7s 103ms/step - loss: 1.5165 - accuracy: 0.4531 - val_loss: 1.3466 - val_accuracy: 0.5125
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.4882 - accuracy: 0.4766
Epoch 72: val_accuracy improved from 0.53125 to 0.53750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 28s 437ms/step - loss: 1.4882 - accuracy: 0.4766 - val_loss: 1.2840 - val_accuracy: 0.5375
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.4849 - accuracy: 0.5029
Epoch 73: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 100ms/step - loss: 1.4849 - accuracy: 0.5029 - val_loss: 1.4033 - val_accuracy: 0.4938
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.7128 - accuracy: 0.3916
Epoch 74: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 102ms/step - loss: 1.7128 - accuracy: 0.3916 - val_loss: 1.3646 - val_accuracy: 0.4875
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.4944 - accuracy: 0.4590
Epoch 75: val_accuracy did not improve from 0.53750
64/64 [==============================] - 6s 101ms/step - loss: 1.4944 - accuracy: 0.4590 - val_loss: 1.5519 - val_accuracy: 0.4187
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.5283 - accuracy: 0.4463
Epoch 76: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 101ms/step - loss: 1.5283 - accuracy: 0.4463 - val_loss: 1.4079 - val_accuracy: 0.4625
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.5194 - accuracy: 0.4316
Epoch 77: val_accuracy improved from 0.53750 to 0.54375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 26s 407ms/step - loss: 1.5194 - accuracy: 0.4316 - val_loss: 1.2427 - val_accuracy: 0.5437
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.4621 - accuracy: 0.4590
Epoch 78: val_accuracy improved from 0.54375 to 0.56250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 29s 456ms/step - loss: 1.4621 - accuracy: 0.4590 - val_loss: 1.2224 - val_accuracy: 0.5625
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.5039 - accuracy: 0.4385
Epoch 79: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 100ms/step - loss: 1.5039 - accuracy: 0.4385 - val_loss: 1.4229 - val_accuracy: 0.4938
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.4822 - accuracy: 0.4639
Epoch 80: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 102ms/step - loss: 1.4822 - accuracy: 0.4639 - val_loss: 1.3038 - val_accuracy: 0.4437
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.4883 - accuracy: 0.4736
Epoch 81: val_accuracy did not improve from 0.56250
64/64 [==============================] - 6s 101ms/step - loss: 1.4883 - accuracy: 0.4736 - val_loss: 1.3777 - val_accuracy: 0.4750
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.4597 - accuracy: 0.4746
Epoch 82: val_accuracy did not improve from 0.56250
64/64 [==============================] - 6s 101ms/step - loss: 1.4597 - accuracy: 0.4746 - val_loss: 1.5361 - val_accuracy: 0.4563
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.4880 - accuracy: 0.4541
Epoch 83: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 102ms/step - loss: 1.4880 - accuracy: 0.4541 - val_loss: 1.2326 - val_accuracy: 0.5625
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.4388 - accuracy: 0.4717
Epoch 84: val_accuracy did not improve from 0.56250
64/64 [==============================] - 6s 101ms/step - loss: 1.4388 - accuracy: 0.4717 - val_loss: 1.3539 - val_accuracy: 0.5000
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.4020 - accuracy: 0.4971
Epoch 85: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 102ms/step - loss: 1.4020 - accuracy: 0.4971 - val_loss: 1.4432 - val_accuracy: 0.4313
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.4528 - accuracy: 0.4824
Epoch 86: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 101ms/step - loss: 1.4528 - accuracy: 0.4824 - val_loss: 1.2468 - val_accuracy: 0.5500
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.5384 - accuracy: 0.4326
Epoch 87: val_accuracy did not improve from 0.56250
64/64 [==============================] - 7s 102ms/step - loss: 1.5384 - accuracy: 0.4326 - val_loss: 1.7168 - val_accuracy: 0.3938
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.4315 - accuracy: 0.4961
Epoch 88: val_accuracy improved from 0.56250 to 0.59375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_3.h5
64/64 [==============================] - 30s 474ms/step - loss: 1.4315 - accuracy: 0.4961 - val_loss: 1.1803 - val_accuracy: 0.5938
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.4566 - accuracy: 0.4932
Epoch 89: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4566 - accuracy: 0.4932 - val_loss: 1.2681 - val_accuracy: 0.5562
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.4521 - accuracy: 0.4775
Epoch 90: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4521 - accuracy: 0.4775 - val_loss: 1.1674 - val_accuracy: 0.5625
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.4657 - accuracy: 0.4697
Epoch 91: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 104ms/step - loss: 1.4657 - accuracy: 0.4697 - val_loss: 1.4733 - val_accuracy: 0.4500
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.4633 - accuracy: 0.4883
Epoch 92: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4633 - accuracy: 0.4883 - val_loss: 1.3237 - val_accuracy: 0.4938
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.4024 - accuracy: 0.5039
Epoch 93: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4024 - accuracy: 0.5039 - val_loss: 1.3462 - val_accuracy: 0.4625
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.4779 - accuracy: 0.4434
Epoch 94: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4779 - accuracy: 0.4434 - val_loss: 1.3641 - val_accuracy: 0.5750
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.4365 - accuracy: 0.4863
Epoch 95: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4365 - accuracy: 0.4863 - val_loss: 1.3506 - val_accuracy: 0.4875
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.4714 - accuracy: 0.4717
Epoch 96: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4714 - accuracy: 0.4717 - val_loss: 1.5053 - val_accuracy: 0.4938
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.4196 - accuracy: 0.4844
Epoch 97: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4196 - accuracy: 0.4844 - val_loss: 1.2417 - val_accuracy: 0.5437
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.5400 - accuracy: 0.4736
Epoch 98: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.5400 - accuracy: 0.4736 - val_loss: 1.2565 - val_accuracy: 0.5250
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.4239 - accuracy: 0.5029
Epoch 99: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4239 - accuracy: 0.5029 - val_loss: 1.1840 - val_accuracy: 0.5688
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.4348 - accuracy: 0.4873
Epoch 100: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4348 - accuracy: 0.4873 - val_loss: 1.1736 - val_accuracy: 0.5813
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.5563 - accuracy: 0.4570
Epoch 101: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.5563 - accuracy: 0.4570 - val_loss: 1.2303 - val_accuracy: 0.5375
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.4324 - accuracy: 0.5068
Epoch 102: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 104ms/step - loss: 1.4324 - accuracy: 0.5068 - val_loss: 1.1423 - val_accuracy: 0.5688
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.4602 - accuracy: 0.4727
Epoch 103: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4602 - accuracy: 0.4727 - val_loss: 1.1249 - val_accuracy: 0.5875
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4334 - accuracy: 0.4863
Epoch 104: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 103ms/step - loss: 1.4334 - accuracy: 0.4863 - val_loss: 1.2156 - val_accuracy: 0.5500
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.4069 - accuracy: 0.5244
Epoch 105: val_accuracy did not improve from 0.59375
64/64 [==============================] - 7s 102ms/step - loss: 1.4069 - accuracy: 0.5244 - val_loss: 1.3548 - val_accuracy: 0.4875


 ********* Training time: 1146.234375 s.


*****************
* Model Summary *
*****************
Model: "model_2"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_3 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense_2 (Dense)             (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 119,595,020
Non-trainable params: 14,714,688
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 25ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.8929    0.5357    0.6696       280
anarrhichomenum     0.4729    0.9028    0.6207       319
   brevantherum     0.5339    0.4719    0.5010       267
      dulcamara     0.2262    0.0603    0.0952       315
   herposolanum     0.4834    0.6300    0.5470       300
     holophylla     0.4184    0.6507    0.5093       335
     lasiocarpa     0.9339    0.6861    0.7910       309
      melongena     0.8099    0.3439    0.4828       285
    micracantha     0.3604    0.3776    0.3688       294
         petota     0.2807    0.4485    0.3453       301
        solanum     0.3456    0.4080    0.3742       299
          torva     0.2871    0.0980    0.1461       296

       accuracy                         0.4714      3600
      macro avg     0.5038    0.4678    0.4543      3600
   weighted avg     0.4994    0.4714    0.4538      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

***************************
* Started at 3531.6875... *
***************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 14.9412 - accuracy: 0.1270
Epoch 1: val_accuracy improved from -inf to 0.12500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 23s 360ms/step - loss: 14.9412 - accuracy: 0.1270 - val_loss: 3.7976 - val_accuracy: 0.1250
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.2030 - accuracy: 0.1904
Epoch 2: val_accuracy improved from 0.12500 to 0.21875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 31s 494ms/step - loss: 2.2030 - accuracy: 0.1904 - val_loss: 1.9452 - val_accuracy: 0.2188
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 1.9537 - accuracy: 0.2549
Epoch 3: val_accuracy improved from 0.21875 to 0.23750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 24s 379ms/step - loss: 1.9537 - accuracy: 0.2549 - val_loss: 2.1034 - val_accuracy: 0.2375
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 1.9347 - accuracy: 0.2793
Epoch 4: val_accuracy improved from 0.23750 to 0.30000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 33s 515ms/step - loss: 1.9347 - accuracy: 0.2793 - val_loss: 1.8453 - val_accuracy: 0.3000
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 1.9421 - accuracy: 0.3066
Epoch 5: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 99ms/step - loss: 1.9421 - accuracy: 0.3066 - val_loss: 1.8907 - val_accuracy: 0.2438
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 1.8948 - accuracy: 0.2939
Epoch 6: val_accuracy improved from 0.30000 to 0.35000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 33s 525ms/step - loss: 1.8948 - accuracy: 0.2939 - val_loss: 1.8866 - val_accuracy: 0.3500
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 1.8345 - accuracy: 0.3145
Epoch 7: val_accuracy improved from 0.35000 to 0.39375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 24s 371ms/step - loss: 1.8345 - accuracy: 0.3145 - val_loss: 1.7512 - val_accuracy: 0.3938
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 1.8595 - accuracy: 0.3105
Epoch 8: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 100ms/step - loss: 1.8595 - accuracy: 0.3105 - val_loss: 2.0200 - val_accuracy: 0.2688
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 1.9035 - accuracy: 0.2764
Epoch 9: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.9035 - accuracy: 0.2764 - val_loss: 1.7339 - val_accuracy: 0.3250
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 1.8584 - accuracy: 0.2979
Epoch 10: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 101ms/step - loss: 1.8584 - accuracy: 0.2979 - val_loss: 1.8556 - val_accuracy: 0.2812
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 1.8014 - accuracy: 0.3193
Epoch 11: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.8014 - accuracy: 0.3193 - val_loss: 1.7100 - val_accuracy: 0.3688
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 1.8038 - accuracy: 0.3125
Epoch 12: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 103ms/step - loss: 1.8038 - accuracy: 0.3125 - val_loss: 1.6928 - val_accuracy: 0.3562
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 1.8258 - accuracy: 0.3066
Epoch 13: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.8258 - accuracy: 0.3066 - val_loss: 1.7335 - val_accuracy: 0.3625
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 1.8078 - accuracy: 0.3262
Epoch 14: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.8078 - accuracy: 0.3262 - val_loss: 1.8149 - val_accuracy: 0.3125
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 1.7854 - accuracy: 0.3389
Epoch 15: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 103ms/step - loss: 1.7854 - accuracy: 0.3389 - val_loss: 1.6382 - val_accuracy: 0.3750
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 1.7081 - accuracy: 0.3584
Epoch 16: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.7081 - accuracy: 0.3584 - val_loss: 1.7223 - val_accuracy: 0.3500
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 1.8104 - accuracy: 0.2812
Epoch 17: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.8104 - accuracy: 0.2812 - val_loss: 1.7064 - val_accuracy: 0.3812
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 1.7065 - accuracy: 0.3652
Epoch 18: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.7065 - accuracy: 0.3652 - val_loss: 1.9505 - val_accuracy: 0.2937
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 1.7679 - accuracy: 0.3271
Epoch 19: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.7679 - accuracy: 0.3271 - val_loss: 1.7534 - val_accuracy: 0.3750
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 1.7466 - accuracy: 0.3516
Epoch 20: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 103ms/step - loss: 1.7466 - accuracy: 0.3516 - val_loss: 1.8553 - val_accuracy: 0.2812
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 1.7133 - accuracy: 0.3379
Epoch 21: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 104ms/step - loss: 1.7133 - accuracy: 0.3379 - val_loss: 1.8979 - val_accuracy: 0.3000
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 1.7385 - accuracy: 0.3525
Epoch 22: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 103ms/step - loss: 1.7385 - accuracy: 0.3525 - val_loss: 1.7378 - val_accuracy: 0.3938
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 1.6840 - accuracy: 0.3672
Epoch 23: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.6840 - accuracy: 0.3672 - val_loss: 1.6105 - val_accuracy: 0.3812
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 1.6164 - accuracy: 0.4023
Epoch 24: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 103ms/step - loss: 1.6164 - accuracy: 0.4023 - val_loss: 1.7361 - val_accuracy: 0.3562
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 1.6394 - accuracy: 0.4092
Epoch 25: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 102ms/step - loss: 1.6394 - accuracy: 0.4092 - val_loss: 1.7402 - val_accuracy: 0.3375
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 1.6748 - accuracy: 0.4004
Epoch 26: val_accuracy improved from 0.39375 to 0.42500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 31s 489ms/step - loss: 1.6748 - accuracy: 0.4004 - val_loss: 1.5788 - val_accuracy: 0.4250
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 1.6914 - accuracy: 0.3818
Epoch 27: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 100ms/step - loss: 1.6914 - accuracy: 0.3818 - val_loss: 1.5882 - val_accuracy: 0.3812
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 1.6692 - accuracy: 0.3926
Epoch 28: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 102ms/step - loss: 1.6692 - accuracy: 0.3926 - val_loss: 1.8413 - val_accuracy: 0.3688
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.6796 - accuracy: 0.3984
Epoch 29: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 103ms/step - loss: 1.6796 - accuracy: 0.3984 - val_loss: 1.8061 - val_accuracy: 0.3250
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.6838 - accuracy: 0.3926
Epoch 30: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 102ms/step - loss: 1.6838 - accuracy: 0.3926 - val_loss: 1.7603 - val_accuracy: 0.3438
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.6427 - accuracy: 0.3936
Epoch 31: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 103ms/step - loss: 1.6427 - accuracy: 0.3936 - val_loss: 1.9952 - val_accuracy: 0.3000
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.6561 - accuracy: 0.4023
Epoch 32: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 103ms/step - loss: 1.6561 - accuracy: 0.4023 - val_loss: 1.6409 - val_accuracy: 0.3750
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.6020 - accuracy: 0.4121
Epoch 33: val_accuracy did not improve from 0.42500
64/64 [==============================] - 7s 103ms/step - loss: 1.6020 - accuracy: 0.4121 - val_loss: 1.7092 - val_accuracy: 0.3562
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.6179 - accuracy: 0.3945
Epoch 34: val_accuracy did not improve from 0.42500
64/64 [==============================] - 6s 101ms/step - loss: 1.6179 - accuracy: 0.3945 - val_loss: 1.7288 - val_accuracy: 0.3875
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.5634 - accuracy: 0.4160
Epoch 35: val_accuracy improved from 0.42500 to 0.43750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 26s 407ms/step - loss: 1.5634 - accuracy: 0.4160 - val_loss: 1.6478 - val_accuracy: 0.4375
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.5896 - accuracy: 0.4160
Epoch 36: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 101ms/step - loss: 1.5896 - accuracy: 0.4160 - val_loss: 1.6849 - val_accuracy: 0.4375
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.5916 - accuracy: 0.4111
Epoch 37: val_accuracy did not improve from 0.43750
64/64 [==============================] - 6s 101ms/step - loss: 1.5916 - accuracy: 0.4111 - val_loss: 1.8327 - val_accuracy: 0.3688
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.5660 - accuracy: 0.4424
Epoch 38: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 102ms/step - loss: 1.5660 - accuracy: 0.4424 - val_loss: 1.5364 - val_accuracy: 0.4375
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.6446 - accuracy: 0.4004
Epoch 39: val_accuracy did not improve from 0.43750
64/64 [==============================] - 6s 101ms/step - loss: 1.6446 - accuracy: 0.4004 - val_loss: 1.8757 - val_accuracy: 0.3438
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.6103 - accuracy: 0.4072
Epoch 40: val_accuracy did not improve from 0.43750
64/64 [==============================] - 6s 101ms/step - loss: 1.6103 - accuracy: 0.4072 - val_loss: 1.6443 - val_accuracy: 0.4062
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.6014 - accuracy: 0.4424
Epoch 41: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 101ms/step - loss: 1.6014 - accuracy: 0.4424 - val_loss: 1.7201 - val_accuracy: 0.3688
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.6151 - accuracy: 0.4180
Epoch 42: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 102ms/step - loss: 1.6151 - accuracy: 0.4180 - val_loss: 1.7790 - val_accuracy: 0.3750
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.5743 - accuracy: 0.4082
Epoch 43: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 101ms/step - loss: 1.5743 - accuracy: 0.4082 - val_loss: 1.6526 - val_accuracy: 0.3938
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.5588 - accuracy: 0.4131
Epoch 44: val_accuracy improved from 0.43750 to 0.47500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 31s 485ms/step - loss: 1.5588 - accuracy: 0.4131 - val_loss: 1.4713 - val_accuracy: 0.4750
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.5304 - accuracy: 0.4326
Epoch 45: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 100ms/step - loss: 1.5304 - accuracy: 0.4326 - val_loss: 1.7106 - val_accuracy: 0.4125
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.5035 - accuracy: 0.4570
Epoch 46: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.5035 - accuracy: 0.4570 - val_loss: 1.5575 - val_accuracy: 0.4250
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.5700 - accuracy: 0.4404
Epoch 47: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 103ms/step - loss: 1.5700 - accuracy: 0.4404 - val_loss: 1.8036 - val_accuracy: 0.4187
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.5844 - accuracy: 0.4463
Epoch 48: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 101ms/step - loss: 1.5844 - accuracy: 0.4463 - val_loss: 1.9296 - val_accuracy: 0.3625
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.4822 - accuracy: 0.4893
Epoch 49: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 103ms/step - loss: 1.4822 - accuracy: 0.4893 - val_loss: 1.5667 - val_accuracy: 0.4625
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.5249 - accuracy: 0.4443
Epoch 50: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.5249 - accuracy: 0.4443 - val_loss: 1.5451 - val_accuracy: 0.4437
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.5694 - accuracy: 0.4541
Epoch 51: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 103ms/step - loss: 1.5694 - accuracy: 0.4541 - val_loss: 1.5625 - val_accuracy: 0.4313
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.4735 - accuracy: 0.4580
Epoch 52: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.4735 - accuracy: 0.4580 - val_loss: 1.6499 - val_accuracy: 0.4062
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.4722 - accuracy: 0.4678
Epoch 53: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 103ms/step - loss: 1.4722 - accuracy: 0.4678 - val_loss: 1.6872 - val_accuracy: 0.4250
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.5548 - accuracy: 0.4453
Epoch 54: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.5548 - accuracy: 0.4453 - val_loss: 1.7391 - val_accuracy: 0.4062
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.5761 - accuracy: 0.4336
Epoch 55: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.5761 - accuracy: 0.4336 - val_loss: 1.5048 - val_accuracy: 0.4437
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.5338 - accuracy: 0.4395
Epoch 56: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 102ms/step - loss: 1.5338 - accuracy: 0.4395 - val_loss: 1.4263 - val_accuracy: 0.4563
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.5591 - accuracy: 0.4492
Epoch 57: val_accuracy did not improve from 0.47500
64/64 [==============================] - 7s 103ms/step - loss: 1.5591 - accuracy: 0.4492 - val_loss: 1.5784 - val_accuracy: 0.4688
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.4914 - accuracy: 0.4834
Epoch 58: val_accuracy improved from 0.47500 to 0.50625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 27s 432ms/step - loss: 1.4914 - accuracy: 0.4834 - val_loss: 1.3772 - val_accuracy: 0.5063
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.5538 - accuracy: 0.4365
Epoch 59: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 101ms/step - loss: 1.5538 - accuracy: 0.4365 - val_loss: 1.5898 - val_accuracy: 0.4375
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.5323 - accuracy: 0.4707
Epoch 60: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5323 - accuracy: 0.4707 - val_loss: 1.5324 - val_accuracy: 0.4750
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.5163 - accuracy: 0.4746
Epoch 61: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.5163 - accuracy: 0.4746 - val_loss: 1.9624 - val_accuracy: 0.3125
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.5434 - accuracy: 0.4541
Epoch 62: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5434 - accuracy: 0.4541 - val_loss: 1.5093 - val_accuracy: 0.4313
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.5212 - accuracy: 0.4512
Epoch 63: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5212 - accuracy: 0.4512 - val_loss: 1.5850 - val_accuracy: 0.3938
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.5042 - accuracy: 0.4590
Epoch 64: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5042 - accuracy: 0.4590 - val_loss: 1.7761 - val_accuracy: 0.4062
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.5283 - accuracy: 0.4629
Epoch 65: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5283 - accuracy: 0.4629 - val_loss: 1.6532 - val_accuracy: 0.3688
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.5118 - accuracy: 0.4648
Epoch 66: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5118 - accuracy: 0.4648 - val_loss: 1.5611 - val_accuracy: 0.4750
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.5653 - accuracy: 0.4541
Epoch 67: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5653 - accuracy: 0.4541 - val_loss: 1.6316 - val_accuracy: 0.4187
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.5038 - accuracy: 0.4531
Epoch 68: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.5038 - accuracy: 0.4531 - val_loss: 1.6310 - val_accuracy: 0.4250
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.5273 - accuracy: 0.4395
Epoch 69: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 101ms/step - loss: 1.5273 - accuracy: 0.4395 - val_loss: 1.4613 - val_accuracy: 0.4563
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.4453 - accuracy: 0.4746
Epoch 70: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4453 - accuracy: 0.4746 - val_loss: 1.5851 - val_accuracy: 0.4938
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.5070 - accuracy: 0.4541
Epoch 71: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5070 - accuracy: 0.4541 - val_loss: 1.5161 - val_accuracy: 0.4563
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.5056 - accuracy: 0.4463
Epoch 72: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5056 - accuracy: 0.4463 - val_loss: 1.8675 - val_accuracy: 0.3438
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.4868 - accuracy: 0.4590
Epoch 73: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4868 - accuracy: 0.4590 - val_loss: 1.5697 - val_accuracy: 0.4375
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.4765 - accuracy: 0.4736
Epoch 74: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4765 - accuracy: 0.4736 - val_loss: 1.4136 - val_accuracy: 0.4563
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.4216 - accuracy: 0.4932
Epoch 75: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4216 - accuracy: 0.4932 - val_loss: 1.5219 - val_accuracy: 0.4625
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.4208 - accuracy: 0.4941
Epoch 76: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4208 - accuracy: 0.4941 - val_loss: 1.6485 - val_accuracy: 0.3750
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.4914 - accuracy: 0.4746
Epoch 77: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4914 - accuracy: 0.4746 - val_loss: 1.4900 - val_accuracy: 0.4563
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.5254 - accuracy: 0.4365
Epoch 78: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.5254 - accuracy: 0.4365 - val_loss: 1.5956 - val_accuracy: 0.4625
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.4961 - accuracy: 0.4570
Epoch 79: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4961 - accuracy: 0.4570 - val_loss: 1.5447 - val_accuracy: 0.4563
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.4784 - accuracy: 0.4766
Epoch 80: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4784 - accuracy: 0.4766 - val_loss: 1.5455 - val_accuracy: 0.4750
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.4465 - accuracy: 0.4736
Epoch 81: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4465 - accuracy: 0.4736 - val_loss: 1.9084 - val_accuracy: 0.3812
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.4637 - accuracy: 0.4834
Epoch 82: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 102ms/step - loss: 1.4637 - accuracy: 0.4834 - val_loss: 1.9312 - val_accuracy: 0.3875
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.4420 - accuracy: 0.4863
Epoch 83: val_accuracy did not improve from 0.50625
64/64 [==============================] - 7s 103ms/step - loss: 1.4420 - accuracy: 0.4863 - val_loss: 1.5690 - val_accuracy: 0.4375
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.4658 - accuracy: 0.4785
Epoch 84: val_accuracy improved from 0.50625 to 0.54375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16_tl_20230217091009_4.h5
64/64 [==============================] - 32s 508ms/step - loss: 1.4658 - accuracy: 0.4785 - val_loss: 1.3468 - val_accuracy: 0.5437
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.4885 - accuracy: 0.4453
Epoch 85: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 101ms/step - loss: 1.4885 - accuracy: 0.4453 - val_loss: 1.5861 - val_accuracy: 0.5000
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.4045 - accuracy: 0.4785
Epoch 86: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4045 - accuracy: 0.4785 - val_loss: 1.4179 - val_accuracy: 0.5125
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.4069 - accuracy: 0.4932
Epoch 87: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4069 - accuracy: 0.4932 - val_loss: 1.5567 - val_accuracy: 0.4625
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.4485 - accuracy: 0.4834
Epoch 88: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4485 - accuracy: 0.4834 - val_loss: 1.6090 - val_accuracy: 0.4875
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.4744 - accuracy: 0.4746
Epoch 89: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 104ms/step - loss: 1.4744 - accuracy: 0.4746 - val_loss: 1.6301 - val_accuracy: 0.4313
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.4212 - accuracy: 0.5039
Epoch 90: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4212 - accuracy: 0.5039 - val_loss: 1.7306 - val_accuracy: 0.4500
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.4667 - accuracy: 0.4756
Epoch 91: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4667 - accuracy: 0.4756 - val_loss: 1.6038 - val_accuracy: 0.4688
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.4245 - accuracy: 0.5059
Epoch 92: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4245 - accuracy: 0.5059 - val_loss: 1.9219 - val_accuracy: 0.3500
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.4790 - accuracy: 0.4658
Epoch 93: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4790 - accuracy: 0.4658 - val_loss: 1.5015 - val_accuracy: 0.4688
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.4809 - accuracy: 0.4980
Epoch 94: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4809 - accuracy: 0.4980 - val_loss: 1.4786 - val_accuracy: 0.5000
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.4275 - accuracy: 0.4844
Epoch 95: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4275 - accuracy: 0.4844 - val_loss: 1.4574 - val_accuracy: 0.5000
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.4023 - accuracy: 0.5059
Epoch 96: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4023 - accuracy: 0.5059 - val_loss: 1.6669 - val_accuracy: 0.4437
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.4498 - accuracy: 0.4756
Epoch 97: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4498 - accuracy: 0.4756 - val_loss: 1.3901 - val_accuracy: 0.5125
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.4808 - accuracy: 0.4668
Epoch 98: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4808 - accuracy: 0.4668 - val_loss: 1.5109 - val_accuracy: 0.4500
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.4471 - accuracy: 0.4775
Epoch 99: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4471 - accuracy: 0.4775 - val_loss: 1.5076 - val_accuracy: 0.4688
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.4420 - accuracy: 0.5029
Epoch 100: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4420 - accuracy: 0.5029 - val_loss: 1.4942 - val_accuracy: 0.4875
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.4988 - accuracy: 0.4766
Epoch 101: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4988 - accuracy: 0.4766 - val_loss: 1.5585 - val_accuracy: 0.4688
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.4229 - accuracy: 0.4736
Epoch 102: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4229 - accuracy: 0.4736 - val_loss: 1.5906 - val_accuracy: 0.4625
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.4151 - accuracy: 0.4961
Epoch 103: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.4151 - accuracy: 0.4961 - val_loss: 1.5588 - val_accuracy: 0.4375
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4594 - accuracy: 0.5039
Epoch 104: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 102ms/step - loss: 1.4594 - accuracy: 0.5039 - val_loss: 1.4123 - val_accuracy: 0.4750
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.3569 - accuracy: 0.5195
Epoch 105: val_accuracy did not improve from 0.54375
64/64 [==============================] - 7s 103ms/step - loss: 1.3569 - accuracy: 0.5195 - val_loss: 1.4070 - val_accuracy: 0.5437


 ********* Training time: 1141.921875 s.


*****************
* Model Summary *
*****************
Model: "model_3"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_4 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense_3 (Dense)             (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 119,595,020
Non-trainable params: 14,714,688
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 6s 25ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.6557    0.8399    0.7364       331
anarrhichomenum     0.6368    0.4671    0.5389       319
   brevantherum     0.6616    0.4199    0.5137       312
      dulcamara     0.3075    0.3822    0.3408       259
   herposolanum     0.5431    0.4615    0.4990       273
     holophylla     0.3770    0.6646    0.4811       316
     lasiocarpa     0.9944    0.5701    0.7247       314
      melongena     0.8281    0.1755    0.2896       302
    micracantha     0.3710    0.4063    0.3879       315
         petota     0.3967    0.6644    0.4968       289
        solanum     0.3516    0.5904    0.4408       271
          torva     0.3714    0.1304    0.1931       299

       accuracy                         0.4844      3600
      macro avg     0.5412    0.4810    0.4702      3600
   weighted avg     0.5483    0.4844    0.4751      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*******************************
* Mean metrics across 4 folds *
*******************************

	0
accuracy	0.475972
acanthophora.precision	0.767927
acanthophora.recall	0.724700
acanthophora.f1-score	0.713952
acanthophora.support	298.250000
anarrhichomenum.precision	0.595728
anarrhichomenum.recall	0.706986
anarrhichomenum.f1-score	0.602344
anarrhichomenum.support	318.750000
brevantherum.precision	0.571459
brevantherum.recall	0.358761
brevantherum.f1-score	0.425595
brevantherum.support	288.000000
dulcamara.precision	0.324495
dulcamara.recall	0.193336
dulcamara.f1-score	0.220247
dulcamara.support	295.500000
herposolanum.precision	0.434159
herposolanum.recall	0.580904
herposolanum.f1-score	0.479326
herposolanum.support	285.000000
holophylla.precision	0.432474
holophylla.recall	0.567413
holophylla.f1-score	0.482154
holophylla.support	304.750000
lasiocarpa.precision	0.904777
lasiocarpa.recall	0.719096
lasiocarpa.f1-score	0.788687
lasiocarpa.support	312.500000
melongena.precision	0.808505
melongena.recall	0.324550
melongena.f1-score	0.429940
melongena.support	298.000000
micracantha.precision	0.448051
micracantha.recall	0.380374
micracantha.f1-score	0.405778
micracantha.support	300.000000
petota.precision	0.341838
petota.recall	0.632698
petota.f1-score	0.426575
petota.support	304.500000
solanum.precision	0.331441
solanum.recall	0.390233
solanum.f1-score	0.347896
solanum.support	294.500000
torva.precision	0.426752
torva.recall	0.090926
torva.f1-score	0.144653
torva.support	300.250000
macro avg.precision	0.532301
macro avg.recall	0.472498
macro avg.f1-score	0.455596
macro avg.support	3600.000000
weighted avg.precision	0.533449
weighted avg.recall	0.475972
weighted avg.f1-score	0.458047
weighted avg.support	3600.000000

CPU times: total: 1h 17min 53s
Wall time: 1h 7min 28s

ResNET50

%%time

kf = ShuffleSplit(n_splits=CV_FOLDS,
                  test_size=VAL_SIZE,
                  random_state=RANDOM_SEED)

split = 1

resnet50_cv_val_pred = []

for train_index, val_index in kf.split(balanced_training_data):

    resnet50_model = tf.keras.applications.ResNet50(
        pooling="avg",
        include_top=True,
        weights=None,
        classes=NUM_CLASSES,
        classifier_activation='softmax'
    )

    # Add an optimizer
    resnet50_model.compile(optimizer=tf.optimizers.Adam(learning_rate=0.003),
                        loss='categorical_crossentropy',
                        metrics=['accuracy'])
    # resnet50_model.compile(optimizer=tf.optimizers.SGD(learning_rate=0.003, momentum=0.9),
    #                        loss='categorical_crossentropy',
    #                        metrics=['accuracy'])

    training_split_data = balanced_training_data.iloc[train_index]
    val_split_data = balanced_training_data.iloc[val_index]

    # Add a progress bar and save checkpoints
    resnet50_callbacks = [
        create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"resnet50_{TRAINING_RUN_ID}_{split}.h5")),
        tf.keras.callbacks.ProgbarLogger(
            count_mode = 'steps',
            stateful_metrics = None
        ),
        tf.keras.callbacks.EarlyStopping(
            monitor='val_accuracy',
            min_delta=0,
            patience=40,
            verbose=1,
            mode='auto',
            restore_best_weights=True
        )
    ]

    fit_params = {
            "x": training_split_data,
            "epochs": 105,
            "callbacks": resnet50_callbacks,
            "validation_data": val_split_data,
            "steps_per_epoch": 64,
            "validation_steps": 10,
    }

    preproc_func = tf.keras.applications.resnet50.preprocess_input
    
    with tf.device(TRAINING_DEVICE_NAME):

        resnet50_training_history = train_model(resnet50_model,
                                                fit_params=fit_params,
                                                preproc_func=preproc_func)
        
        _, pred_report = evaluate_model(resnet50_model,
                                        resnet50_training_history,
                                        fit_params=fit_params,
                                        preproc_func=preproc_func)
    
        resnet50_cv_val_pred.append(pred_report)
    
    split += 1

calculate_cv_mean_metrics(resnet50_cv_val_pred)

**************************
* Started at 11.59375... *
**************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 3.9523 - accuracy: 0.0850
Epoch 1: val_accuracy improved from -inf to 0.06875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 16s 141ms/step - loss: 3.9523 - accuracy: 0.0850 - val_loss: 3304.0222 - val_accuracy: 0.0688
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.5358 - accuracy: 0.1260
Epoch 2: val_accuracy improved from 0.06875 to 0.13750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 13s 200ms/step - loss: 2.5358 - accuracy: 0.1260 - val_loss: 13.0022 - val_accuracy: 0.1375
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.6443 - accuracy: 0.1436
Epoch 3: val_accuracy did not improve from 0.13750
64/64 [==============================] - 13s 196ms/step - loss: 2.6443 - accuracy: 0.1436 - val_loss: 361.7343 - val_accuracy: 0.0625
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 2.4638 - accuracy: 0.1494
Epoch 4: val_accuracy did not improve from 0.13750
64/64 [==============================] - 8s 129ms/step - loss: 2.4638 - accuracy: 0.1494 - val_loss: 2.4240 - val_accuracy: 0.1063
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 2.3624 - accuracy: 0.1768
Epoch 5: val_accuracy improved from 0.13750 to 0.15625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 10s 153ms/step - loss: 2.3624 - accuracy: 0.1768 - val_loss: 2.4630 - val_accuracy: 0.1562
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.3109 - accuracy: 0.2041
Epoch 6: val_accuracy improved from 0.15625 to 0.20000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 11s 174ms/step - loss: 2.3109 - accuracy: 0.2041 - val_loss: 2.3138 - val_accuracy: 0.2000
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.3030 - accuracy: 0.1895
Epoch 7: val_accuracy did not improve from 0.20000
64/64 [==============================] - 8s 131ms/step - loss: 2.3030 - accuracy: 0.1895 - val_loss: 25.8597 - val_accuracy: 0.1562
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.2849 - accuracy: 0.1943
Epoch 8: val_accuracy did not improve from 0.20000
64/64 [==============================] - 8s 121ms/step - loss: 2.2849 - accuracy: 0.1943 - val_loss: 5.4640 - val_accuracy: 0.1375
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.2538 - accuracy: 0.1895
Epoch 9: val_accuracy did not improve from 0.20000
64/64 [==============================] - 8s 121ms/step - loss: 2.2538 - accuracy: 0.1895 - val_loss: 5.8237 - val_accuracy: 0.1000
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.1991 - accuracy: 0.2090
Epoch 10: val_accuracy improved from 0.20000 to 0.21875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 8s 131ms/step - loss: 2.1991 - accuracy: 0.2090 - val_loss: 2.6069 - val_accuracy: 0.2188
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.2279 - accuracy: 0.2109
Epoch 11: val_accuracy did not improve from 0.21875
64/64 [==============================] - 9s 131ms/step - loss: 2.2279 - accuracy: 0.2109 - val_loss: 3.3335 - val_accuracy: 0.1875
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.1907 - accuracy: 0.2412
Epoch 12: val_accuracy did not improve from 0.21875
64/64 [==============================] - 8s 117ms/step - loss: 2.1907 - accuracy: 0.2412 - val_loss: 4.5206 - val_accuracy: 0.1250
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.1448 - accuracy: 0.2363
Epoch 13: val_accuracy improved from 0.21875 to 0.23750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 142ms/step - loss: 2.1448 - accuracy: 0.2363 - val_loss: 2.2018 - val_accuracy: 0.2375
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.1114 - accuracy: 0.2490
Epoch 14: val_accuracy did not improve from 0.23750
64/64 [==============================] - 8s 124ms/step - loss: 2.1114 - accuracy: 0.2490 - val_loss: 2.5284 - val_accuracy: 0.2313
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.0689 - accuracy: 0.2598
Epoch 15: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 115ms/step - loss: 2.0689 - accuracy: 0.2598 - val_loss: 3.0121 - val_accuracy: 0.1625
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.0941 - accuracy: 0.2363
Epoch 16: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 114ms/step - loss: 2.0941 - accuracy: 0.2363 - val_loss: 3.8181 - val_accuracy: 0.2000
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.0380 - accuracy: 0.2607
Epoch 17: val_accuracy improved from 0.23750 to 0.25625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 133ms/step - loss: 2.0380 - accuracy: 0.2607 - val_loss: 2.4819 - val_accuracy: 0.2562
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.0299 - accuracy: 0.2666
Epoch 18: val_accuracy did not improve from 0.25625
64/64 [==============================] - 8s 117ms/step - loss: 2.0299 - accuracy: 0.2666 - val_loss: 2.3683 - val_accuracy: 0.2250
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.0635 - accuracy: 0.2549
Epoch 19: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 112ms/step - loss: 2.0635 - accuracy: 0.2549 - val_loss: 2.9197 - val_accuracy: 0.1562
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.0713 - accuracy: 0.2607
Epoch 20: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 113ms/step - loss: 2.0713 - accuracy: 0.2607 - val_loss: 7.3077 - val_accuracy: 0.1250
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.0474 - accuracy: 0.2812
Epoch 21: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 116ms/step - loss: 2.0474 - accuracy: 0.2812 - val_loss: 2.2546 - val_accuracy: 0.2125
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.0342 - accuracy: 0.2695
Epoch 22: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 114ms/step - loss: 2.0342 - accuracy: 0.2695 - val_loss: 10.2068 - val_accuracy: 0.1063
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.0281 - accuracy: 0.2793
Epoch 23: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 116ms/step - loss: 2.0281 - accuracy: 0.2793 - val_loss: 5.0967 - val_accuracy: 0.2062
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.0272 - accuracy: 0.2744
Epoch 24: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 114ms/step - loss: 2.0272 - accuracy: 0.2744 - val_loss: 3.9561 - val_accuracy: 0.1688
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 1.9860 - accuracy: 0.2979
Epoch 25: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 114ms/step - loss: 1.9860 - accuracy: 0.2979 - val_loss: 4.1490 - val_accuracy: 0.1625
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 1.9542 - accuracy: 0.2979
Epoch 26: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 114ms/step - loss: 1.9542 - accuracy: 0.2979 - val_loss: 3.5106 - val_accuracy: 0.2062
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.0085 - accuracy: 0.2783
Epoch 27: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 112ms/step - loss: 2.0085 - accuracy: 0.2783 - val_loss: 5.1532 - val_accuracy: 0.1625
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 1.9019 - accuracy: 0.3096
Epoch 28: val_accuracy improved from 0.25625 to 0.30625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 136ms/step - loss: 1.9019 - accuracy: 0.3096 - val_loss: 1.9293 - val_accuracy: 0.3063
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.9803 - accuracy: 0.2939
Epoch 29: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.9803 - accuracy: 0.2939 - val_loss: 3.3186 - val_accuracy: 0.1500
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.9498 - accuracy: 0.3105
Epoch 30: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.9498 - accuracy: 0.3105 - val_loss: 2.0330 - val_accuracy: 0.2625
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.9434 - accuracy: 0.3008
Epoch 31: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 1.9434 - accuracy: 0.3008 - val_loss: 2.4606 - val_accuracy: 0.2562
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.9491 - accuracy: 0.3154
Epoch 32: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 114ms/step - loss: 1.9491 - accuracy: 0.3154 - val_loss: 2.9831 - val_accuracy: 0.2500
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.8720 - accuracy: 0.3066
Epoch 33: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.8720 - accuracy: 0.3066 - val_loss: 3.2116 - val_accuracy: 0.2625
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.9014 - accuracy: 0.3330
Epoch 34: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.9014 - accuracy: 0.3330 - val_loss: 2.5069 - val_accuracy: 0.2062
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 2.0149 - accuracy: 0.2832
Epoch 35: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.0149 - accuracy: 0.2832 - val_loss: 38.3172 - val_accuracy: 0.1562
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.9780 - accuracy: 0.2988
Epoch 36: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.9780 - accuracy: 0.2988 - val_loss: 2.9295 - val_accuracy: 0.2000
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.9236 - accuracy: 0.3076
Epoch 37: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 1.9236 - accuracy: 0.3076 - val_loss: 6.0876 - val_accuracy: 0.2313
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.9806 - accuracy: 0.2773
Epoch 38: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.9806 - accuracy: 0.2773 - val_loss: 2.2363 - val_accuracy: 0.2250
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.9538 - accuracy: 0.2979
Epoch 39: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.9538 - accuracy: 0.2979 - val_loss: 6.6918 - val_accuracy: 0.1312
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.8993 - accuracy: 0.3076
Epoch 40: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.8993 - accuracy: 0.3076 - val_loss: 9.1859 - val_accuracy: 0.1312
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.9046 - accuracy: 0.3184
Epoch 41: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.9046 - accuracy: 0.3184 - val_loss: 2.4718 - val_accuracy: 0.2688
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.9266 - accuracy: 0.3232
Epoch 42: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 1.9266 - accuracy: 0.3232 - val_loss: 3.0511 - val_accuracy: 0.2562
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.9161 - accuracy: 0.3174
Epoch 43: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.9161 - accuracy: 0.3174 - val_loss: 4.8519 - val_accuracy: 0.1875
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.8713 - accuracy: 0.3359
Epoch 44: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.8713 - accuracy: 0.3359 - val_loss: 5.5246 - val_accuracy: 0.2062
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.9069 - accuracy: 0.3281
Epoch 45: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.9069 - accuracy: 0.3281 - val_loss: 2.0925 - val_accuracy: 0.2937
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.8806 - accuracy: 0.3389
Epoch 46: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 1.8806 - accuracy: 0.3389 - val_loss: 4.3142 - val_accuracy: 0.2500
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.8739 - accuracy: 0.3311
Epoch 47: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 1.8739 - accuracy: 0.3311 - val_loss: 2.1895 - val_accuracy: 0.2688
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.8454 - accuracy: 0.3301
Epoch 48: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 1.8454 - accuracy: 0.3301 - val_loss: 2.3883 - val_accuracy: 0.3063
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.8319 - accuracy: 0.3496
Epoch 49: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 1.8319 - accuracy: 0.3496 - val_loss: 2.4354 - val_accuracy: 0.2562
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.8577 - accuracy: 0.3477
Epoch 50: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 109ms/step - loss: 1.8577 - accuracy: 0.3477 - val_loss: 2.2787 - val_accuracy: 0.2875
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.8529 - accuracy: 0.3613
Epoch 51: val_accuracy improved from 0.30625 to 0.33750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 8s 126ms/step - loss: 1.8529 - accuracy: 0.3613 - val_loss: 1.9150 - val_accuracy: 0.3375
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.8803 - accuracy: 0.3379
Epoch 52: val_accuracy did not improve from 0.33750
64/64 [==============================] - 7s 110ms/step - loss: 1.8803 - accuracy: 0.3379 - val_loss: 2.0774 - val_accuracy: 0.2750
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.8852 - accuracy: 0.3242
Epoch 53: val_accuracy improved from 0.33750 to 0.34375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 8s 124ms/step - loss: 1.8852 - accuracy: 0.3242 - val_loss: 1.9119 - val_accuracy: 0.3438
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.8527 - accuracy: 0.3438
Epoch 54: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 109ms/step - loss: 1.8527 - accuracy: 0.3438 - val_loss: 2.9542 - val_accuracy: 0.2500
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.7882 - accuracy: 0.3428
Epoch 55: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 109ms/step - loss: 1.7882 - accuracy: 0.3428 - val_loss: 3.0582 - val_accuracy: 0.2625
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.8778 - accuracy: 0.3350
Epoch 56: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 109ms/step - loss: 1.8778 - accuracy: 0.3350 - val_loss: 2.2699 - val_accuracy: 0.3000
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.8786 - accuracy: 0.3154
Epoch 57: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 109ms/step - loss: 1.8786 - accuracy: 0.3154 - val_loss: 2.2414 - val_accuracy: 0.2875
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.8158 - accuracy: 0.3311
Epoch 58: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 112ms/step - loss: 1.8158 - accuracy: 0.3311 - val_loss: 2.9257 - val_accuracy: 0.2438
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.7976 - accuracy: 0.3457
Epoch 59: val_accuracy improved from 0.34375 to 0.40625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 8s 128ms/step - loss: 1.7976 - accuracy: 0.3457 - val_loss: 1.8131 - val_accuracy: 0.4062
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.8129 - accuracy: 0.3428
Epoch 60: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 1.8129 - accuracy: 0.3428 - val_loss: 3.0188 - val_accuracy: 0.2750
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.8813 - accuracy: 0.3369
Epoch 61: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.8813 - accuracy: 0.3369 - val_loss: 5.3609 - val_accuracy: 0.1937
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.8767 - accuracy: 0.3311
Epoch 62: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.8767 - accuracy: 0.3311 - val_loss: 5.5975 - val_accuracy: 0.1437
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.8553 - accuracy: 0.3506
Epoch 63: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 109ms/step - loss: 1.8553 - accuracy: 0.3506 - val_loss: 6.0375 - val_accuracy: 0.1750
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.7814 - accuracy: 0.3770
Epoch 64: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 1.7814 - accuracy: 0.3770 - val_loss: 2.0218 - val_accuracy: 0.3562
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.8135 - accuracy: 0.3682
Epoch 65: val_accuracy improved from 0.40625 to 0.41250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 134ms/step - loss: 1.8135 - accuracy: 0.3682 - val_loss: 2.0649 - val_accuracy: 0.4125
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.7086 - accuracy: 0.3896
Epoch 66: val_accuracy did not improve from 0.41250
64/64 [==============================] - 7s 111ms/step - loss: 1.7086 - accuracy: 0.3896 - val_loss: 2.2259 - val_accuracy: 0.4000
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.7595 - accuracy: 0.3662
Epoch 67: val_accuracy improved from 0.41250 to 0.43750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 142ms/step - loss: 1.7595 - accuracy: 0.3662 - val_loss: 1.8778 - val_accuracy: 0.4375
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.7371 - accuracy: 0.3877
Epoch 68: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7371 - accuracy: 0.3877 - val_loss: 1.8521 - val_accuracy: 0.4000
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.7651 - accuracy: 0.3672
Epoch 69: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7651 - accuracy: 0.3672 - val_loss: 1.8758 - val_accuracy: 0.3313
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.7485 - accuracy: 0.3770
Epoch 70: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7485 - accuracy: 0.3770 - val_loss: 2.0379 - val_accuracy: 0.3313
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.7851 - accuracy: 0.3877
Epoch 71: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7851 - accuracy: 0.3877 - val_loss: 6.0165 - val_accuracy: 0.2062
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.7607 - accuracy: 0.3838
Epoch 72: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7607 - accuracy: 0.3838 - val_loss: 2.1400 - val_accuracy: 0.3375
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.7333 - accuracy: 0.3545
Epoch 73: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 113ms/step - loss: 1.7333 - accuracy: 0.3545 - val_loss: 1.7137 - val_accuracy: 0.3688
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.7759 - accuracy: 0.3838
Epoch 74: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 115ms/step - loss: 1.7759 - accuracy: 0.3838 - val_loss: 2.3283 - val_accuracy: 0.3375
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.7496 - accuracy: 0.3828
Epoch 75: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7496 - accuracy: 0.3828 - val_loss: 1.7013 - val_accuracy: 0.4062
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.7037 - accuracy: 0.4336
Epoch 76: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7037 - accuracy: 0.4336 - val_loss: 2.0851 - val_accuracy: 0.3438
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.7343 - accuracy: 0.3936
Epoch 77: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 113ms/step - loss: 1.7343 - accuracy: 0.3936 - val_loss: 2.0351 - val_accuracy: 0.3063
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.7121 - accuracy: 0.3877
Epoch 78: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7121 - accuracy: 0.3877 - val_loss: 4.1976 - val_accuracy: 0.2625
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.7005 - accuracy: 0.3867
Epoch 79: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7005 - accuracy: 0.3867 - val_loss: 12.1819 - val_accuracy: 0.1500
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.6815 - accuracy: 0.4160
Epoch 80: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 110ms/step - loss: 1.6815 - accuracy: 0.4160 - val_loss: 2.7275 - val_accuracy: 0.2875
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.7073 - accuracy: 0.4023
Epoch 81: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7073 - accuracy: 0.4023 - val_loss: 2.3153 - val_accuracy: 0.3375
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.7259 - accuracy: 0.3945
Epoch 82: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.7259 - accuracy: 0.3945 - val_loss: 1.8843 - val_accuracy: 0.4250
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.8228 - accuracy: 0.3662
Epoch 83: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 114ms/step - loss: 1.8228 - accuracy: 0.3662 - val_loss: 7.8654 - val_accuracy: 0.2625
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.7581 - accuracy: 0.3730
Epoch 84: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.7581 - accuracy: 0.3730 - val_loss: 2.6081 - val_accuracy: 0.3562
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.6946 - accuracy: 0.4189
Epoch 85: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.6946 - accuracy: 0.4189 - val_loss: 2.3701 - val_accuracy: 0.3313
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.6484 - accuracy: 0.4238
Epoch 86: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.6484 - accuracy: 0.4238 - val_loss: 2.4030 - val_accuracy: 0.3250
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.5804 - accuracy: 0.4365
Epoch 87: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 114ms/step - loss: 1.5804 - accuracy: 0.4365 - val_loss: 1.8291 - val_accuracy: 0.4313
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.6702 - accuracy: 0.3984
Epoch 88: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 112ms/step - loss: 1.6702 - accuracy: 0.3984 - val_loss: 2.0837 - val_accuracy: 0.3125
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.7060 - accuracy: 0.3926
Epoch 89: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 114ms/step - loss: 1.7060 - accuracy: 0.3926 - val_loss: 3.6748 - val_accuracy: 0.1813
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.6419 - accuracy: 0.4111
Epoch 90: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 111ms/step - loss: 1.6419 - accuracy: 0.4111 - val_loss: 2.0762 - val_accuracy: 0.3438
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.6250 - accuracy: 0.4248
Epoch 91: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 114ms/step - loss: 1.6250 - accuracy: 0.4248 - val_loss: 2.0953 - val_accuracy: 0.3438
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.5839 - accuracy: 0.4307
Epoch 92: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 110ms/step - loss: 1.5839 - accuracy: 0.4307 - val_loss: 3.0040 - val_accuracy: 0.2125
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.6052 - accuracy: 0.4346
Epoch 93: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 114ms/step - loss: 1.6052 - accuracy: 0.4346 - val_loss: 2.1340 - val_accuracy: 0.3750
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.6357 - accuracy: 0.4297
Epoch 94: val_accuracy improved from 0.43750 to 0.48125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_1.h5
64/64 [==============================] - 9s 136ms/step - loss: 1.6357 - accuracy: 0.4297 - val_loss: 2.1239 - val_accuracy: 0.4812
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.5992 - accuracy: 0.4365
Epoch 95: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 109ms/step - loss: 1.5992 - accuracy: 0.4365 - val_loss: 3.5609 - val_accuracy: 0.2750
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.5889 - accuracy: 0.4297
Epoch 96: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 110ms/step - loss: 1.5889 - accuracy: 0.4297 - val_loss: 4.5117 - val_accuracy: 0.2812
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.6302 - accuracy: 0.4268
Epoch 97: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 109ms/step - loss: 1.6302 - accuracy: 0.4268 - val_loss: 2.0663 - val_accuracy: 0.3812
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.5556 - accuracy: 0.4482
Epoch 98: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 109ms/step - loss: 1.5556 - accuracy: 0.4482 - val_loss: 2.8550 - val_accuracy: 0.2313
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.5884 - accuracy: 0.4434
Epoch 99: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 110ms/step - loss: 1.5884 - accuracy: 0.4434 - val_loss: 3.0940 - val_accuracy: 0.3875
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.6178 - accuracy: 0.4395
Epoch 100: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 110ms/step - loss: 1.6178 - accuracy: 0.4395 - val_loss: 1.9225 - val_accuracy: 0.3812
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.5507 - accuracy: 0.4600
Epoch 101: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 116ms/step - loss: 1.5507 - accuracy: 0.4600 - val_loss: 1.8372 - val_accuracy: 0.4125
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.4970 - accuracy: 0.4727
Epoch 102: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 108ms/step - loss: 1.4970 - accuracy: 0.4727 - val_loss: 2.1633 - val_accuracy: 0.3375
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.5522 - accuracy: 0.4473
Epoch 103: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 104ms/step - loss: 1.5522 - accuracy: 0.4473 - val_loss: 2.9998 - val_accuracy: 0.2188
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4440 - accuracy: 0.4756
Epoch 104: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 104ms/step - loss: 1.4440 - accuracy: 0.4756 - val_loss: 3.6900 - val_accuracy: 0.2625
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.5333 - accuracy: 0.4590
Epoch 105: val_accuracy did not improve from 0.48125
64/64 [==============================] - 7s 103ms/step - loss: 1.5333 - accuracy: 0.4590 - val_loss: 2.4637 - val_accuracy: 0.3625


 ********* Training time: 1431.578125 s.


*****************
* Model Summary *
*****************
Model: "resnet50"
__________________________________________________________________________________________________
 Layer (type)                   Output Shape         Param #     Connected to                     
==================================================================================================
 input_1 (InputLayer)           [(None, 224, 224, 3  0           []                               
                                )]                                                                
                                                                                                  
 conv1_pad (ZeroPadding2D)      (None, 230, 230, 3)  0           ['input_1[0][0]']                
                                                                                                  
 conv1_conv (Conv2D)            (None, 112, 112, 64  9472        ['conv1_pad[0][0]']              
                                )                                                                 
                                                                                                  
 conv1_bn (BatchNormalization)  (None, 112, 112, 64  256         ['conv1_conv[0][0]']             
                                )                                                                 
                                                                                                  
 conv1_relu (Activation)        (None, 112, 112, 64  0           ['conv1_bn[0][0]']               
                                )                                                                 
                                                                                                  
 pool1_pad (ZeroPadding2D)      (None, 114, 114, 64  0           ['conv1_relu[0][0]']             
                                )                                                                 
                                                                                                  
 pool1_pool (MaxPooling2D)      (None, 56, 56, 64)   0           ['pool1_pad[0][0]']              
                                                                                                  
 conv2_block1_1_conv (Conv2D)   (None, 56, 56, 64)   4160        ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block1_1_relu[0][0]']    
                                                                                                  
 conv2_block1_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_0_conv (Conv2D)   (None, 56, 56, 256)  16640       ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block1_2_relu[0][0]']    
                                                                                                  
 conv2_block1_0_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_0_bn[0][0]',      
                                                                  'conv2_block1_3_bn[0][0]']      
                                                                                                  
 conv2_block1_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block1_add[0][0]']       
                                                                                                  
 conv2_block2_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block1_out[0][0]']       
                                                                                                  
 conv2_block2_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block2_1_relu[0][0]']    
                                                                                                  
 conv2_block2_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block2_2_relu[0][0]']    
                                                                                                  
 conv2_block2_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_out[0][0]',       
                                                                  'conv2_block2_3_bn[0][0]']      
                                                                                                  
 conv2_block2_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block2_add[0][0]']       
                                                                                                  
 conv2_block3_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block2_out[0][0]']       
                                                                                                  
 conv2_block3_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block3_1_relu[0][0]']    
                                                                                                  
 conv2_block3_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block3_2_relu[0][0]']    
                                                                                                  
 conv2_block3_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_add (Add)         (None, 56, 56, 256)  0           ['conv2_block2_out[0][0]',       
                                                                  'conv2_block3_3_bn[0][0]']      
                                                                                                  
 conv2_block3_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block3_add[0][0]']       
                                                                                                  
 conv3_block1_1_conv (Conv2D)   (None, 28, 28, 128)  32896       ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block1_1_relu[0][0]']    
                                                                                                  
 conv3_block1_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_0_conv (Conv2D)   (None, 28, 28, 512)  131584      ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block1_2_relu[0][0]']    
                                                                                                  
 conv3_block1_0_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_0_bn[0][0]',      
                                                                  'conv3_block1_3_bn[0][0]']      
                                                                                                  
 conv3_block1_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block1_add[0][0]']       
                                                                                                  
 conv3_block2_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block1_out[0][0]']       
                                                                                                  
 conv3_block2_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block2_1_relu[0][0]']    
                                                                                                  
 conv3_block2_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block2_2_relu[0][0]']    
                                                                                                  
 conv3_block2_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_out[0][0]',       
                                                                  'conv3_block2_3_bn[0][0]']      
                                                                                                  
 conv3_block2_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block2_add[0][0]']       
                                                                                                  
 conv3_block3_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block2_out[0][0]']       
                                                                                                  
 conv3_block3_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block3_1_relu[0][0]']    
                                                                                                  
 conv3_block3_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block3_2_relu[0][0]']    
                                                                                                  
 conv3_block3_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_add (Add)         (None, 28, 28, 512)  0           ['conv3_block2_out[0][0]',       
                                                                  'conv3_block3_3_bn[0][0]']      
                                                                                                  
 conv3_block3_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block3_add[0][0]']       
                                                                                                  
 conv3_block4_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block3_out[0][0]']       
                                                                                                  
 conv3_block4_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block4_1_relu[0][0]']    
                                                                                                  
 conv3_block4_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block4_2_relu[0][0]']    
                                                                                                  
 conv3_block4_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block4_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_add (Add)         (None, 28, 28, 512)  0           ['conv3_block3_out[0][0]',       
                                                                  'conv3_block4_3_bn[0][0]']      
                                                                                                  
 conv3_block4_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block4_add[0][0]']       
                                                                                                  
 conv4_block1_1_conv (Conv2D)   (None, 14, 14, 256)  131328      ['conv3_block4_out[0][0]']       
                                                                                                  
 conv4_block1_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block1_1_relu[0][0]']    
                                                                                                  
 conv4_block1_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_0_conv (Conv2D)   (None, 14, 14, 1024  525312      ['conv3_block4_out[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block1_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block1_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block1_0_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_0_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_0_bn[0][0]',      
                                )                                 'conv4_block1_3_bn[0][0]']      
                                                                                                  
 conv4_block1_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block1_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block2_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block1_out[0][0]']       
                                                                                                  
 conv4_block2_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block2_1_relu[0][0]']    
                                                                                                  
 conv4_block2_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block2_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block2_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block2_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block2_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_out[0][0]',       
                                )                                 'conv4_block2_3_bn[0][0]']      
                                                                                                  
 conv4_block2_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block2_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block3_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block2_out[0][0]']       
                                                                                                  
 conv4_block3_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block3_1_relu[0][0]']    
                                                                                                  
 conv4_block3_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block3_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block3_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block3_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block3_add (Add)         (None, 14, 14, 1024  0           ['conv4_block2_out[0][0]',       
                                )                                 'conv4_block3_3_bn[0][0]']      
                                                                                                  
 conv4_block3_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block3_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block4_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block3_out[0][0]']       
                                                                                                  
 conv4_block4_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block4_1_relu[0][0]']    
                                                                                                  
 conv4_block4_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block4_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block4_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block4_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block4_add (Add)         (None, 14, 14, 1024  0           ['conv4_block3_out[0][0]',       
                                )                                 'conv4_block4_3_bn[0][0]']      
                                                                                                  
 conv4_block4_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block4_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block5_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block4_out[0][0]']       
                                                                                                  
 conv4_block5_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block5_1_relu[0][0]']    
                                                                                                  
 conv4_block5_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block5_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block5_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block5_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block5_add (Add)         (None, 14, 14, 1024  0           ['conv4_block4_out[0][0]',       
                                )                                 'conv4_block5_3_bn[0][0]']      
                                                                                                  
 conv4_block5_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block5_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block6_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block5_out[0][0]']       
                                                                                                  
 conv4_block6_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block6_1_relu[0][0]']    
                                                                                                  
 conv4_block6_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block6_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block6_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block6_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block6_add (Add)         (None, 14, 14, 1024  0           ['conv4_block5_out[0][0]',       
                                )                                 'conv4_block6_3_bn[0][0]']      
                                                                                                  
 conv4_block6_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block6_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv5_block1_1_conv (Conv2D)   (None, 7, 7, 512)    524800      ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block1_1_relu[0][0]']    
                                                                                                  
 conv5_block1_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_0_conv (Conv2D)   (None, 7, 7, 2048)   2099200     ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block1_2_relu[0][0]']    
                                                                                                  
 conv5_block1_0_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_0_bn[0][0]',      
                                                                  'conv5_block1_3_bn[0][0]']      
                                                                                                  
 conv5_block1_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block1_add[0][0]']       
                                                                                                  
 conv5_block2_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block1_out[0][0]']       
                                                                                                  
 conv5_block2_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block2_1_relu[0][0]']    
                                                                                                  
 conv5_block2_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block2_2_relu[0][0]']    
                                                                                                  
 conv5_block2_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_out[0][0]',       
                                                                  'conv5_block2_3_bn[0][0]']      
                                                                                                  
 conv5_block2_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block2_add[0][0]']       
                                                                                                  
 conv5_block3_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block2_out[0][0]']       
                                                                                                  
 conv5_block3_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block3_1_relu[0][0]']    
                                                                                                  
 conv5_block3_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block3_2_relu[0][0]']    
                                                                                                  
 conv5_block3_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block2_out[0][0]',       
                                                                  'conv5_block3_3_bn[0][0]']      
                                                                                                  
 conv5_block3_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block3_add[0][0]']       
                                                                                                  
 avg_pool (GlobalAveragePooling  (None, 2048)        0           ['conv5_block3_out[0][0]']       
 2D)                                                                                              
                                                                                                  
 predictions (Dense)            (None, 12)           24588       ['avg_pool[0][0]']               
                                                                                                  
==================================================================================================
Total params: 23,612,300
Trainable params: 23,559,180
Non-trainable params: 53,120
__________________________________________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 29ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.6504    0.4820    0.5537       305
anarrhichomenum     0.5552    0.5932    0.5736       322
   brevantherum     0.2778    0.3195    0.2972       266
      dulcamara     0.4500    0.0289    0.0544       311
   herposolanum     0.1684    0.4767    0.2488       279
     holophylla     0.6028    0.3014    0.4019       282
     lasiocarpa     0.6964    0.5495    0.6143       313
      melongena     0.3437    0.3885    0.3647       314
    micracantha     0.2135    0.3813    0.2737       299
         petota     0.2681    0.4700    0.3414       300
        solanum     0.4595    0.0559    0.0997       304
          torva     0.3784    0.0918    0.1478       305

       accuracy                         0.3456      3600
      macro avg     0.4220    0.3449    0.3309      3600
   weighted avg     0.4259    0.3456    0.3330      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*****************************
* Started at 1455.921875... *
*****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 3.9492 - accuracy: 0.0811
Epoch 1: val_accuracy improved from -inf to 0.08125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 12s 123ms/step - loss: 3.9492 - accuracy: 0.0811 - val_loss: 14135.5830 - val_accuracy: 0.0812
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.8075 - accuracy: 0.1084
Epoch 2: val_accuracy did not improve from 0.08125
64/64 [==============================] - 7s 103ms/step - loss: 2.8075 - accuracy: 0.1084 - val_loss: 44.9530 - val_accuracy: 0.0812
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.5349 - accuracy: 0.1465
Epoch 3: val_accuracy did not improve from 0.08125
64/64 [==============================] - 7s 103ms/step - loss: 2.5349 - accuracy: 0.1465 - val_loss: 3.3257 - val_accuracy: 0.0688
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 2.4418 - accuracy: 0.1338
Epoch 4: val_accuracy improved from 0.08125 to 0.10000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 126ms/step - loss: 2.4418 - accuracy: 0.1338 - val_loss: 2.4350 - val_accuracy: 0.1000
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 2.3709 - accuracy: 0.1318
Epoch 5: val_accuracy improved from 0.10000 to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 130ms/step - loss: 2.3709 - accuracy: 0.1318 - val_loss: 2.4182 - val_accuracy: 0.1063
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.3991 - accuracy: 0.1543
Epoch 6: val_accuracy improved from 0.10625 to 0.11875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 126ms/step - loss: 2.3991 - accuracy: 0.1543 - val_loss: 7.9676 - val_accuracy: 0.1187
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.3467 - accuracy: 0.1875
Epoch 7: val_accuracy improved from 0.11875 to 0.14375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 121ms/step - loss: 2.3467 - accuracy: 0.1875 - val_loss: 2.7334 - val_accuracy: 0.1437
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.3497 - accuracy: 0.1582
Epoch 8: val_accuracy did not improve from 0.14375
64/64 [==============================] - 7s 103ms/step - loss: 2.3497 - accuracy: 0.1582 - val_loss: 2.7379 - val_accuracy: 0.1187
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.3157 - accuracy: 0.1641
Epoch 9: val_accuracy improved from 0.14375 to 0.17500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 130ms/step - loss: 2.3157 - accuracy: 0.1641 - val_loss: 2.3707 - val_accuracy: 0.1750
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.2658 - accuracy: 0.1875
Epoch 10: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 103ms/step - loss: 2.2658 - accuracy: 0.1875 - val_loss: 3.3253 - val_accuracy: 0.0938
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.2826 - accuracy: 0.1787
Epoch 11: val_accuracy improved from 0.17500 to 0.20000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.2826 - accuracy: 0.1787 - val_loss: 2.6546 - val_accuracy: 0.2000
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.2375 - accuracy: 0.1973
Epoch 12: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.2375 - accuracy: 0.1973 - val_loss: 2.7180 - val_accuracy: 0.1875
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.1769 - accuracy: 0.2051
Epoch 13: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1769 - accuracy: 0.2051 - val_loss: 4.4300 - val_accuracy: 0.1125
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.1638 - accuracy: 0.2266
Epoch 14: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1638 - accuracy: 0.2266 - val_loss: 4.3419 - val_accuracy: 0.1000
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.2292 - accuracy: 0.1934
Epoch 15: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.2292 - accuracy: 0.1934 - val_loss: 2.3835 - val_accuracy: 0.1937
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.1595 - accuracy: 0.2178
Epoch 16: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1595 - accuracy: 0.2178 - val_loss: 2.4833 - val_accuracy: 0.1312
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.2145 - accuracy: 0.2393
Epoch 17: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 102ms/step - loss: 2.2145 - accuracy: 0.2393 - val_loss: 2.5687 - val_accuracy: 0.1375
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.1662 - accuracy: 0.2334
Epoch 18: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1662 - accuracy: 0.2334 - val_loss: 7.8775 - val_accuracy: 0.1000
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.1583 - accuracy: 0.2148
Epoch 19: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1583 - accuracy: 0.2148 - val_loss: 2.8118 - val_accuracy: 0.1562
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.1106 - accuracy: 0.2354
Epoch 20: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 103ms/step - loss: 2.1106 - accuracy: 0.2354 - val_loss: 8.0629 - val_accuracy: 0.1250
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.1213 - accuracy: 0.2559
Epoch 21: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 102ms/step - loss: 2.1213 - accuracy: 0.2559 - val_loss: 2.5138 - val_accuracy: 0.1813
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.0591 - accuracy: 0.2666
Epoch 22: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 102ms/step - loss: 2.0591 - accuracy: 0.2666 - val_loss: 2.5504 - val_accuracy: 0.1875
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.0989 - accuracy: 0.2568
Epoch 23: val_accuracy did not improve from 0.20000
64/64 [==============================] - 7s 102ms/step - loss: 2.0989 - accuracy: 0.2568 - val_loss: 4.8717 - val_accuracy: 0.0938
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.1386 - accuracy: 0.2354
Epoch 24: val_accuracy improved from 0.20000 to 0.21250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 127ms/step - loss: 2.1386 - accuracy: 0.2354 - val_loss: 2.2663 - val_accuracy: 0.2125
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.0665 - accuracy: 0.2686
Epoch 25: val_accuracy improved from 0.21250 to 0.22500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 121ms/step - loss: 2.0665 - accuracy: 0.2686 - val_loss: 3.0575 - val_accuracy: 0.2250
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.1153 - accuracy: 0.2363
Epoch 26: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.1153 - accuracy: 0.2363 - val_loss: 3.3487 - val_accuracy: 0.2062
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.0457 - accuracy: 0.2539
Epoch 27: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.0457 - accuracy: 0.2539 - val_loss: 2.5023 - val_accuracy: 0.2125
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.0508 - accuracy: 0.2695
Epoch 28: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.0508 - accuracy: 0.2695 - val_loss: 3.3212 - val_accuracy: 0.2000
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.0522 - accuracy: 0.2812
Epoch 29: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.0522 - accuracy: 0.2812 - val_loss: 3.6510 - val_accuracy: 0.1500
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 1.9626 - accuracy: 0.2832
Epoch 30: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 108ms/step - loss: 1.9626 - accuracy: 0.2832 - val_loss: 2.9248 - val_accuracy: 0.1937
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.9844 - accuracy: 0.2783
Epoch 31: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 1.9844 - accuracy: 0.2783 - val_loss: 2.7748 - val_accuracy: 0.2250
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 1.9672 - accuracy: 0.3096
Epoch 32: val_accuracy improved from 0.22500 to 0.29375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 121ms/step - loss: 1.9672 - accuracy: 0.3096 - val_loss: 2.2173 - val_accuracy: 0.2937
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.9647 - accuracy: 0.2861
Epoch 33: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9647 - accuracy: 0.2861 - val_loss: 2.8646 - val_accuracy: 0.1688
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.9664 - accuracy: 0.2793
Epoch 34: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9664 - accuracy: 0.2793 - val_loss: 2.8372 - val_accuracy: 0.2250
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.9486 - accuracy: 0.2939
Epoch 35: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 102ms/step - loss: 1.9486 - accuracy: 0.2939 - val_loss: 2.5567 - val_accuracy: 0.2375
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.9454 - accuracy: 0.2988
Epoch 36: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 104ms/step - loss: 1.9454 - accuracy: 0.2988 - val_loss: 3.9853 - val_accuracy: 0.1187
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.9873 - accuracy: 0.3086
Epoch 37: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 104ms/step - loss: 1.9873 - accuracy: 0.3086 - val_loss: 3.4425 - val_accuracy: 0.1500
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.9810 - accuracy: 0.2910
Epoch 38: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 102ms/step - loss: 1.9810 - accuracy: 0.2910 - val_loss: 2.9893 - val_accuracy: 0.1750
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.9651 - accuracy: 0.3135
Epoch 39: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9651 - accuracy: 0.3135 - val_loss: 2.2778 - val_accuracy: 0.2313
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.9322 - accuracy: 0.3154
Epoch 40: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 102ms/step - loss: 1.9322 - accuracy: 0.3154 - val_loss: 4.9351 - val_accuracy: 0.1187
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.9389 - accuracy: 0.3105
Epoch 41: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9389 - accuracy: 0.3105 - val_loss: 3.1086 - val_accuracy: 0.1875
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.9809 - accuracy: 0.2998
Epoch 42: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 104ms/step - loss: 1.9809 - accuracy: 0.2998 - val_loss: 2.0711 - val_accuracy: 0.2562
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.8847 - accuracy: 0.3135
Epoch 43: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 102ms/step - loss: 1.8847 - accuracy: 0.3135 - val_loss: 2.5209 - val_accuracy: 0.2500
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.9466 - accuracy: 0.2930
Epoch 44: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9466 - accuracy: 0.2930 - val_loss: 2.3347 - val_accuracy: 0.2750
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.9199 - accuracy: 0.2969
Epoch 45: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 108ms/step - loss: 1.9199 - accuracy: 0.2969 - val_loss: 8.2957 - val_accuracy: 0.1063
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.9653 - accuracy: 0.2793
Epoch 46: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 103ms/step - loss: 1.9653 - accuracy: 0.2793 - val_loss: 2.7098 - val_accuracy: 0.1437
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.9521 - accuracy: 0.3076
Epoch 47: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 109ms/step - loss: 1.9521 - accuracy: 0.3076 - val_loss: 3.6050 - val_accuracy: 0.1500
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.9922 - accuracy: 0.2803
Epoch 48: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 108ms/step - loss: 1.9922 - accuracy: 0.2803 - val_loss: 2.4093 - val_accuracy: 0.2250
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.8879 - accuracy: 0.3223
Epoch 49: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 112ms/step - loss: 1.8879 - accuracy: 0.3223 - val_loss: 2.2718 - val_accuracy: 0.2188
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.9060 - accuracy: 0.3350
Epoch 50: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 109ms/step - loss: 1.9060 - accuracy: 0.3350 - val_loss: 2.1596 - val_accuracy: 0.1937
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.9327 - accuracy: 0.3135
Epoch 51: val_accuracy did not improve from 0.29375
64/64 [==============================] - 7s 109ms/step - loss: 1.9327 - accuracy: 0.3135 - val_loss: 2.4265 - val_accuracy: 0.2812
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.8341 - accuracy: 0.3486
Epoch 52: val_accuracy improved from 0.29375 to 0.34375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 129ms/step - loss: 1.8341 - accuracy: 0.3486 - val_loss: 2.0512 - val_accuracy: 0.3438
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.9426 - accuracy: 0.3105
Epoch 53: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 112ms/step - loss: 1.9426 - accuracy: 0.3105 - val_loss: 14.7166 - val_accuracy: 0.0812
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.9492 - accuracy: 0.3203
Epoch 54: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 109ms/step - loss: 1.9492 - accuracy: 0.3203 - val_loss: 2.2813 - val_accuracy: 0.3063
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.8767 - accuracy: 0.3164
Epoch 55: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8767 - accuracy: 0.3164 - val_loss: 5.5869 - val_accuracy: 0.1500
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.8768 - accuracy: 0.3203
Epoch 56: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 111ms/step - loss: 1.8768 - accuracy: 0.3203 - val_loss: 2.4572 - val_accuracy: 0.2812
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.8180 - accuracy: 0.3506
Epoch 57: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 115ms/step - loss: 1.8180 - accuracy: 0.3506 - val_loss: 3.6784 - val_accuracy: 0.1813
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.8380 - accuracy: 0.3398
Epoch 58: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 112ms/step - loss: 1.8380 - accuracy: 0.3398 - val_loss: 2.2510 - val_accuracy: 0.2500
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.7936 - accuracy: 0.3604
Epoch 59: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 112ms/step - loss: 1.7936 - accuracy: 0.3604 - val_loss: 1.9983 - val_accuracy: 0.3125
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.8365 - accuracy: 0.3359
Epoch 60: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8365 - accuracy: 0.3359 - val_loss: 7.2656 - val_accuracy: 0.1500
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.8503 - accuracy: 0.3516
Epoch 61: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 116ms/step - loss: 1.8503 - accuracy: 0.3516 - val_loss: 2.4111 - val_accuracy: 0.2250
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.8371 - accuracy: 0.3594
Epoch 62: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 113ms/step - loss: 1.8371 - accuracy: 0.3594 - val_loss: 2.0011 - val_accuracy: 0.3000
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.7545 - accuracy: 0.3730
Epoch 63: val_accuracy did not improve from 0.34375
64/64 [==============================] - 8s 118ms/step - loss: 1.7545 - accuracy: 0.3730 - val_loss: 2.0879 - val_accuracy: 0.3063
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.8539 - accuracy: 0.3398
Epoch 64: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 111ms/step - loss: 1.8539 - accuracy: 0.3398 - val_loss: 2.3298 - val_accuracy: 0.2812
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.8061 - accuracy: 0.3262
Epoch 65: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8061 - accuracy: 0.3262 - val_loss: 1.8843 - val_accuracy: 0.3313
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.8332 - accuracy: 0.3623
Epoch 66: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8332 - accuracy: 0.3623 - val_loss: 1.9289 - val_accuracy: 0.3187
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.8059 - accuracy: 0.3604
Epoch 67: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8059 - accuracy: 0.3604 - val_loss: 2.0810 - val_accuracy: 0.3063
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.7866 - accuracy: 0.3652
Epoch 68: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 111ms/step - loss: 1.7866 - accuracy: 0.3652 - val_loss: 2.2435 - val_accuracy: 0.2875
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.8386 - accuracy: 0.3418
Epoch 69: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8386 - accuracy: 0.3418 - val_loss: 2.5762 - val_accuracy: 0.2125
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.8262 - accuracy: 0.3525
Epoch 70: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.8262 - accuracy: 0.3525 - val_loss: 2.9027 - val_accuracy: 0.2062
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.7733 - accuracy: 0.3916
Epoch 71: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 110ms/step - loss: 1.7733 - accuracy: 0.3916 - val_loss: 2.2186 - val_accuracy: 0.2812
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.7191 - accuracy: 0.3838
Epoch 72: val_accuracy improved from 0.34375 to 0.35000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 9s 138ms/step - loss: 1.7191 - accuracy: 0.3838 - val_loss: 1.8722 - val_accuracy: 0.3500
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.7893 - accuracy: 0.3779
Epoch 73: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7893 - accuracy: 0.3779 - val_loss: 1.9282 - val_accuracy: 0.2812
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.7682 - accuracy: 0.3643
Epoch 74: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7682 - accuracy: 0.3643 - val_loss: 2.6997 - val_accuracy: 0.2000
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.7050 - accuracy: 0.4033
Epoch 75: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7050 - accuracy: 0.4033 - val_loss: 3.1208 - val_accuracy: 0.2375
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.7500 - accuracy: 0.3809
Epoch 76: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7500 - accuracy: 0.3809 - val_loss: 4.3332 - val_accuracy: 0.1562
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.7377 - accuracy: 0.3828
Epoch 77: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7377 - accuracy: 0.3828 - val_loss: 2.6843 - val_accuracy: 0.2500
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.7189 - accuracy: 0.4092
Epoch 78: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 116ms/step - loss: 1.7189 - accuracy: 0.4092 - val_loss: 1.9775 - val_accuracy: 0.3125
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.7388 - accuracy: 0.3994
Epoch 79: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.7388 - accuracy: 0.3994 - val_loss: 2.8636 - val_accuracy: 0.1937
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.6868 - accuracy: 0.3838
Epoch 80: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 107ms/step - loss: 1.6868 - accuracy: 0.3838 - val_loss: 4.6486 - val_accuracy: 0.1875
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.7512 - accuracy: 0.3926
Epoch 81: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 102ms/step - loss: 1.7512 - accuracy: 0.3926 - val_loss: 5.6962 - val_accuracy: 0.1500
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.7561 - accuracy: 0.3779
Epoch 82: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7561 - accuracy: 0.3779 - val_loss: 2.9087 - val_accuracy: 0.2125
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.6631 - accuracy: 0.4307
Epoch 83: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 102ms/step - loss: 1.6631 - accuracy: 0.4307 - val_loss: 2.8781 - val_accuracy: 0.2625
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.6681 - accuracy: 0.4131
Epoch 84: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6681 - accuracy: 0.4131 - val_loss: 1.9893 - val_accuracy: 0.3250
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.6827 - accuracy: 0.4033
Epoch 85: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6827 - accuracy: 0.4033 - val_loss: 2.2343 - val_accuracy: 0.3438
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.6781 - accuracy: 0.4141
Epoch 86: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6781 - accuracy: 0.4141 - val_loss: 2.7768 - val_accuracy: 0.2375
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.6345 - accuracy: 0.4316
Epoch 87: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6345 - accuracy: 0.4316 - val_loss: 2.3754 - val_accuracy: 0.2188
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.5694 - accuracy: 0.4434
Epoch 88: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.5694 - accuracy: 0.4434 - val_loss: 7.8269 - val_accuracy: 0.1562
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.6148 - accuracy: 0.4307
Epoch 89: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6148 - accuracy: 0.4307 - val_loss: 2.0808 - val_accuracy: 0.3187
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.5805 - accuracy: 0.4395
Epoch 90: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 104ms/step - loss: 1.5805 - accuracy: 0.4395 - val_loss: 2.3865 - val_accuracy: 0.2062
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.6039 - accuracy: 0.4453
Epoch 91: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6039 - accuracy: 0.4453 - val_loss: 2.0231 - val_accuracy: 0.3438
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.5993 - accuracy: 0.4502
Epoch 92: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 102ms/step - loss: 1.5993 - accuracy: 0.4502 - val_loss: 2.0590 - val_accuracy: 0.2875
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.6277 - accuracy: 0.4355
Epoch 93: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 103ms/step - loss: 1.6277 - accuracy: 0.4355 - val_loss: 2.6306 - val_accuracy: 0.2812
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.5953 - accuracy: 0.4307
Epoch 94: val_accuracy improved from 0.35000 to 0.37500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 122ms/step - loss: 1.5953 - accuracy: 0.4307 - val_loss: 1.8147 - val_accuracy: 0.3750
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.5531 - accuracy: 0.4639
Epoch 95: val_accuracy did not improve from 0.37500
64/64 [==============================] - 7s 103ms/step - loss: 1.5531 - accuracy: 0.4639 - val_loss: 5.4752 - val_accuracy: 0.1562
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.5612 - accuracy: 0.4375
Epoch 96: val_accuracy did not improve from 0.37500
64/64 [==============================] - 7s 103ms/step - loss: 1.5612 - accuracy: 0.4375 - val_loss: 4.5717 - val_accuracy: 0.2125
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.5974 - accuracy: 0.4326
Epoch 97: val_accuracy improved from 0.37500 to 0.43750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 9s 139ms/step - loss: 1.5974 - accuracy: 0.4326 - val_loss: 1.5994 - val_accuracy: 0.4375
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.5399 - accuracy: 0.4541
Epoch 98: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 104ms/step - loss: 1.5399 - accuracy: 0.4541 - val_loss: 1.9371 - val_accuracy: 0.3500
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.5440 - accuracy: 0.4609
Epoch 99: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 103ms/step - loss: 1.5440 - accuracy: 0.4609 - val_loss: 1.9682 - val_accuracy: 0.3438
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.4973 - accuracy: 0.4902
Epoch 100: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 102ms/step - loss: 1.4973 - accuracy: 0.4902 - val_loss: 2.1751 - val_accuracy: 0.3562
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.5338 - accuracy: 0.4678
Epoch 101: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 103ms/step - loss: 1.5338 - accuracy: 0.4678 - val_loss: 2.1382 - val_accuracy: 0.3500
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.5528 - accuracy: 0.4551
Epoch 102: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 103ms/step - loss: 1.5528 - accuracy: 0.4551 - val_loss: 4.1755 - val_accuracy: 0.1875
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.4915 - accuracy: 0.5020
Epoch 103: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 104ms/step - loss: 1.4915 - accuracy: 0.5020 - val_loss: 2.2785 - val_accuracy: 0.3125
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4208 - accuracy: 0.5166
Epoch 104: val_accuracy did not improve from 0.43750
64/64 [==============================] - 7s 102ms/step - loss: 1.4208 - accuracy: 0.5166 - val_loss: 1.7211 - val_accuracy: 0.4250
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.4726 - accuracy: 0.4873
Epoch 105: val_accuracy improved from 0.43750 to 0.48750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_2.h5
64/64 [==============================] - 8s 121ms/step - loss: 1.4726 - accuracy: 0.4873 - val_loss: 1.6797 - val_accuracy: 0.4875


 ********* Training time: 1402.40625 s.


*****************
* Model Summary *
*****************
Model: "resnet50"
__________________________________________________________________________________________________
 Layer (type)                   Output Shape         Param #     Connected to                     
==================================================================================================
 input_2 (InputLayer)           [(None, 224, 224, 3  0           []                               
                                )]                                                                
                                                                                                  
 conv1_pad (ZeroPadding2D)      (None, 230, 230, 3)  0           ['input_2[0][0]']                
                                                                                                  
 conv1_conv (Conv2D)            (None, 112, 112, 64  9472        ['conv1_pad[0][0]']              
                                )                                                                 
                                                                                                  
 conv1_bn (BatchNormalization)  (None, 112, 112, 64  256         ['conv1_conv[0][0]']             
                                )                                                                 
                                                                                                  
 conv1_relu (Activation)        (None, 112, 112, 64  0           ['conv1_bn[0][0]']               
                                )                                                                 
                                                                                                  
 pool1_pad (ZeroPadding2D)      (None, 114, 114, 64  0           ['conv1_relu[0][0]']             
                                )                                                                 
                                                                                                  
 pool1_pool (MaxPooling2D)      (None, 56, 56, 64)   0           ['pool1_pad[0][0]']              
                                                                                                  
 conv2_block1_1_conv (Conv2D)   (None, 56, 56, 64)   4160        ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block1_1_relu[0][0]']    
                                                                                                  
 conv2_block1_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_0_conv (Conv2D)   (None, 56, 56, 256)  16640       ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block1_2_relu[0][0]']    
                                                                                                  
 conv2_block1_0_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_0_bn[0][0]',      
                                                                  'conv2_block1_3_bn[0][0]']      
                                                                                                  
 conv2_block1_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block1_add[0][0]']       
                                                                                                  
 conv2_block2_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block1_out[0][0]']       
                                                                                                  
 conv2_block2_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block2_1_relu[0][0]']    
                                                                                                  
 conv2_block2_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block2_2_relu[0][0]']    
                                                                                                  
 conv2_block2_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_out[0][0]',       
                                                                  'conv2_block2_3_bn[0][0]']      
                                                                                                  
 conv2_block2_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block2_add[0][0]']       
                                                                                                  
 conv2_block3_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block2_out[0][0]']       
                                                                                                  
 conv2_block3_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block3_1_relu[0][0]']    
                                                                                                  
 conv2_block3_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block3_2_relu[0][0]']    
                                                                                                  
 conv2_block3_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_add (Add)         (None, 56, 56, 256)  0           ['conv2_block2_out[0][0]',       
                                                                  'conv2_block3_3_bn[0][0]']      
                                                                                                  
 conv2_block3_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block3_add[0][0]']       
                                                                                                  
 conv3_block1_1_conv (Conv2D)   (None, 28, 28, 128)  32896       ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block1_1_relu[0][0]']    
                                                                                                  
 conv3_block1_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_0_conv (Conv2D)   (None, 28, 28, 512)  131584      ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block1_2_relu[0][0]']    
                                                                                                  
 conv3_block1_0_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_0_bn[0][0]',      
                                                                  'conv3_block1_3_bn[0][0]']      
                                                                                                  
 conv3_block1_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block1_add[0][0]']       
                                                                                                  
 conv3_block2_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block1_out[0][0]']       
                                                                                                  
 conv3_block2_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block2_1_relu[0][0]']    
                                                                                                  
 conv3_block2_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block2_2_relu[0][0]']    
                                                                                                  
 conv3_block2_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_out[0][0]',       
                                                                  'conv3_block2_3_bn[0][0]']      
                                                                                                  
 conv3_block2_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block2_add[0][0]']       
                                                                                                  
 conv3_block3_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block2_out[0][0]']       
                                                                                                  
 conv3_block3_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block3_1_relu[0][0]']    
                                                                                                  
 conv3_block3_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block3_2_relu[0][0]']    
                                                                                                  
 conv3_block3_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_add (Add)         (None, 28, 28, 512)  0           ['conv3_block2_out[0][0]',       
                                                                  'conv3_block3_3_bn[0][0]']      
                                                                                                  
 conv3_block3_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block3_add[0][0]']       
                                                                                                  
 conv3_block4_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block3_out[0][0]']       
                                                                                                  
 conv3_block4_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block4_1_relu[0][0]']    
                                                                                                  
 conv3_block4_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block4_2_relu[0][0]']    
                                                                                                  
 conv3_block4_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block4_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_add (Add)         (None, 28, 28, 512)  0           ['conv3_block3_out[0][0]',       
                                                                  'conv3_block4_3_bn[0][0]']      
                                                                                                  
 conv3_block4_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block4_add[0][0]']       
                                                                                                  
 conv4_block1_1_conv (Conv2D)   (None, 14, 14, 256)  131328      ['conv3_block4_out[0][0]']       
                                                                                                  
 conv4_block1_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block1_1_relu[0][0]']    
                                                                                                  
 conv4_block1_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_0_conv (Conv2D)   (None, 14, 14, 1024  525312      ['conv3_block4_out[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block1_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block1_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block1_0_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_0_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_0_bn[0][0]',      
                                )                                 'conv4_block1_3_bn[0][0]']      
                                                                                                  
 conv4_block1_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block1_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block2_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block1_out[0][0]']       
                                                                                                  
 conv4_block2_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block2_1_relu[0][0]']    
                                                                                                  
 conv4_block2_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block2_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block2_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block2_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block2_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_out[0][0]',       
                                )                                 'conv4_block2_3_bn[0][0]']      
                                                                                                  
 conv4_block2_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block2_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block3_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block2_out[0][0]']       
                                                                                                  
 conv4_block3_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block3_1_relu[0][0]']    
                                                                                                  
 conv4_block3_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block3_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block3_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block3_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block3_add (Add)         (None, 14, 14, 1024  0           ['conv4_block2_out[0][0]',       
                                )                                 'conv4_block3_3_bn[0][0]']      
                                                                                                  
 conv4_block3_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block3_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block4_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block3_out[0][0]']       
                                                                                                  
 conv4_block4_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block4_1_relu[0][0]']    
                                                                                                  
 conv4_block4_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block4_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block4_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block4_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block4_add (Add)         (None, 14, 14, 1024  0           ['conv4_block3_out[0][0]',       
                                )                                 'conv4_block4_3_bn[0][0]']      
                                                                                                  
 conv4_block4_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block4_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block5_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block4_out[0][0]']       
                                                                                                  
 conv4_block5_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block5_1_relu[0][0]']    
                                                                                                  
 conv4_block5_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block5_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block5_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block5_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block5_add (Add)         (None, 14, 14, 1024  0           ['conv4_block4_out[0][0]',       
                                )                                 'conv4_block5_3_bn[0][0]']      
                                                                                                  
 conv4_block5_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block5_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block6_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block5_out[0][0]']       
                                                                                                  
 conv4_block6_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block6_1_relu[0][0]']    
                                                                                                  
 conv4_block6_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block6_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block6_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block6_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block6_add (Add)         (None, 14, 14, 1024  0           ['conv4_block5_out[0][0]',       
                                )                                 'conv4_block6_3_bn[0][0]']      
                                                                                                  
 conv4_block6_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block6_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv5_block1_1_conv (Conv2D)   (None, 7, 7, 512)    524800      ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block1_1_relu[0][0]']    
                                                                                                  
 conv5_block1_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_0_conv (Conv2D)   (None, 7, 7, 2048)   2099200     ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block1_2_relu[0][0]']    
                                                                                                  
 conv5_block1_0_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_0_bn[0][0]',      
                                                                  'conv5_block1_3_bn[0][0]']      
                                                                                                  
 conv5_block1_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block1_add[0][0]']       
                                                                                                  
 conv5_block2_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block1_out[0][0]']       
                                                                                                  
 conv5_block2_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block2_1_relu[0][0]']    
                                                                                                  
 conv5_block2_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block2_2_relu[0][0]']    
                                                                                                  
 conv5_block2_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_out[0][0]',       
                                                                  'conv5_block2_3_bn[0][0]']      
                                                                                                  
 conv5_block2_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block2_add[0][0]']       
                                                                                                  
 conv5_block3_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block2_out[0][0]']       
                                                                                                  
 conv5_block3_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block3_1_relu[0][0]']    
                                                                                                  
 conv5_block3_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block3_2_relu[0][0]']    
                                                                                                  
 conv5_block3_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block2_out[0][0]',       
                                                                  'conv5_block3_3_bn[0][0]']      
                                                                                                  
 conv5_block3_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block3_add[0][0]']       
                                                                                                  
 avg_pool (GlobalAveragePooling  (None, 2048)        0           ['conv5_block3_out[0][0]']       
 2D)                                                                                              
                                                                                                  
 predictions (Dense)            (None, 12)           24588       ['avg_pool[0][0]']               
                                                                                                  
==================================================================================================
Total params: 23,612,300
Trainable params: 23,559,180
Non-trainable params: 53,120
__________________________________________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 27ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.7705    0.6787    0.7217       277
anarrhichomenum     0.6864    0.5143    0.5880       315
   brevantherum     0.4089    0.6580    0.5044       307
      dulcamara     0.5161    0.3232    0.3975       297
   herposolanum     0.3236    0.5799    0.4154       288
     holophylla     0.5069    0.5140    0.5104       286
     lasiocarpa     0.7052    0.9140    0.7961       314
      melongena     0.8045    0.4948    0.6128       291
    micracantha     0.3454    0.4247    0.3810       292
         petota     0.4078    0.5122    0.4541       328
        solanum     0.6809    0.1053    0.1823       304
          torva     0.3087    0.2359    0.2674       301

       accuracy                         0.4967      3600
      macro avg     0.5387    0.4962    0.4859      3600
   weighted avg     0.5380    0.4967    0.4856      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*****************************
* Started at 2871.140625... *
*****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 4.0296 - accuracy: 0.0947
Epoch 1: val_accuracy improved from -inf to 0.09375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 11s 122ms/step - loss: 4.0296 - accuracy: 0.0947 - val_loss: 7.1974 - val_accuracy: 0.0938
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 3.0052 - accuracy: 0.1094
Epoch 2: val_accuracy did not improve from 0.09375
64/64 [==============================] - 7s 104ms/step - loss: 3.0052 - accuracy: 0.1094 - val_loss: 3251.5769 - val_accuracy: 0.0812
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.6046 - accuracy: 0.1270
Epoch 3: val_accuracy improved from 0.09375 to 0.16875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 120ms/step - loss: 2.6046 - accuracy: 0.1270 - val_loss: 3.1643 - val_accuracy: 0.1688
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 2.4387 - accuracy: 0.1455
Epoch 4: val_accuracy did not improve from 0.16875
64/64 [==============================] - 7s 104ms/step - loss: 2.4387 - accuracy: 0.1455 - val_loss: 3.5246 - val_accuracy: 0.0688
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 2.4326 - accuracy: 0.1709
Epoch 5: val_accuracy did not improve from 0.16875
64/64 [==============================] - 7s 104ms/step - loss: 2.4326 - accuracy: 0.1709 - val_loss: 5.4712 - val_accuracy: 0.1312
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.3068 - accuracy: 0.1885
Epoch 6: val_accuracy improved from 0.16875 to 0.20625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 126ms/step - loss: 2.3068 - accuracy: 0.1885 - val_loss: 3.7744 - val_accuracy: 0.2062
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.3320 - accuracy: 0.1924
Epoch 7: val_accuracy improved from 0.20625 to 0.22500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 128ms/step - loss: 2.3320 - accuracy: 0.1924 - val_loss: 8.3325 - val_accuracy: 0.2250
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.3480 - accuracy: 0.1807
Epoch 8: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.3480 - accuracy: 0.1807 - val_loss: 9.1093 - val_accuracy: 0.1187
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.2883 - accuracy: 0.2100
Epoch 9: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.2883 - accuracy: 0.2100 - val_loss: 2.4915 - val_accuracy: 0.2125
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.2440 - accuracy: 0.1924
Epoch 10: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.2440 - accuracy: 0.1924 - val_loss: 3.9342 - val_accuracy: 0.1937
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.2775 - accuracy: 0.1826
Epoch 11: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.2775 - accuracy: 0.1826 - val_loss: 4.1680 - val_accuracy: 0.1625
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.2751 - accuracy: 0.2109
Epoch 12: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 105ms/step - loss: 2.2751 - accuracy: 0.2109 - val_loss: 2.7190 - val_accuracy: 0.2125
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.2426 - accuracy: 0.2227
Epoch 13: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.2426 - accuracy: 0.2227 - val_loss: 11.9522 - val_accuracy: 0.1875
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.2032 - accuracy: 0.2305
Epoch 14: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.2032 - accuracy: 0.2305 - val_loss: 9.6028 - val_accuracy: 0.2125
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.2771 - accuracy: 0.1836
Epoch 15: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.2771 - accuracy: 0.1836 - val_loss: 3.8325 - val_accuracy: 0.1688
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.2090 - accuracy: 0.2217
Epoch 16: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.2090 - accuracy: 0.2217 - val_loss: 2.6072 - val_accuracy: 0.1875
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.1642 - accuracy: 0.2119
Epoch 17: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.1642 - accuracy: 0.2119 - val_loss: 4.5153 - val_accuracy: 0.2062
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.1629 - accuracy: 0.2314
Epoch 18: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 104ms/step - loss: 2.1629 - accuracy: 0.2314 - val_loss: 2.3892 - val_accuracy: 0.1937
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.1991 - accuracy: 0.2217
Epoch 19: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.1991 - accuracy: 0.2217 - val_loss: 3.9922 - val_accuracy: 0.1750
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.1257 - accuracy: 0.2334
Epoch 20: val_accuracy did not improve from 0.22500
64/64 [==============================] - 7s 103ms/step - loss: 2.1257 - accuracy: 0.2334 - val_loss: 4.5742 - val_accuracy: 0.1875
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.1139 - accuracy: 0.2539
Epoch 21: val_accuracy improved from 0.22500 to 0.23750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 124ms/step - loss: 2.1139 - accuracy: 0.2539 - val_loss: 2.4622 - val_accuracy: 0.2375
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.1390 - accuracy: 0.2344
Epoch 22: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 104ms/step - loss: 2.1390 - accuracy: 0.2344 - val_loss: 2.6034 - val_accuracy: 0.1625
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.1608 - accuracy: 0.2236
Epoch 23: val_accuracy improved from 0.23750 to 0.25000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.1608 - accuracy: 0.2236 - val_loss: 7.1883 - val_accuracy: 0.2500
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.1087 - accuracy: 0.2529
Epoch 24: val_accuracy did not improve from 0.25000
64/64 [==============================] - 7s 104ms/step - loss: 2.1087 - accuracy: 0.2529 - val_loss: 4.0482 - val_accuracy: 0.2438
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.0726 - accuracy: 0.2480
Epoch 25: val_accuracy did not improve from 0.25000
64/64 [==============================] - 7s 104ms/step - loss: 2.0726 - accuracy: 0.2480 - val_loss: 6.9413 - val_accuracy: 0.1500
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.0893 - accuracy: 0.2598
Epoch 26: val_accuracy did not improve from 0.25000
64/64 [==============================] - 7s 104ms/step - loss: 2.0893 - accuracy: 0.2598 - val_loss: 3.5752 - val_accuracy: 0.1312
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.0684 - accuracy: 0.2559
Epoch 27: val_accuracy improved from 0.25000 to 0.28125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 127ms/step - loss: 2.0684 - accuracy: 0.2559 - val_loss: 1.9530 - val_accuracy: 0.2812
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.0238 - accuracy: 0.2754
Epoch 28: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 104ms/step - loss: 2.0238 - accuracy: 0.2754 - val_loss: 2.4389 - val_accuracy: 0.2562
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.0457 - accuracy: 0.2646
Epoch 29: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 103ms/step - loss: 2.0457 - accuracy: 0.2646 - val_loss: 5.8522 - val_accuracy: 0.1063
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.0350 - accuracy: 0.2686
Epoch 30: val_accuracy improved from 0.28125 to 0.28750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 119ms/step - loss: 2.0350 - accuracy: 0.2686 - val_loss: 2.0168 - val_accuracy: 0.2875
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.9661 - accuracy: 0.2900
Epoch 31: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.9661 - accuracy: 0.2900 - val_loss: 2.9514 - val_accuracy: 0.2625
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.0070 - accuracy: 0.2715
Epoch 32: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 2.0070 - accuracy: 0.2715 - val_loss: 6.3692 - val_accuracy: 0.2875
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.9707 - accuracy: 0.2988
Epoch 33: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 104ms/step - loss: 1.9707 - accuracy: 0.2988 - val_loss: 1.9048 - val_accuracy: 0.2562
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 1.9882 - accuracy: 0.2861
Epoch 34: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 104ms/step - loss: 1.9882 - accuracy: 0.2861 - val_loss: 1.9168 - val_accuracy: 0.2812
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.9550 - accuracy: 0.3135
Epoch 35: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.9550 - accuracy: 0.3135 - val_loss: 2.3261 - val_accuracy: 0.2438
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.9450 - accuracy: 0.3271
Epoch 36: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 104ms/step - loss: 1.9450 - accuracy: 0.3271 - val_loss: 2.1259 - val_accuracy: 0.2875
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 1.9726 - accuracy: 0.3008
Epoch 37: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.9726 - accuracy: 0.3008 - val_loss: 2.2057 - val_accuracy: 0.2313
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.9732 - accuracy: 0.2822
Epoch 38: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.9732 - accuracy: 0.2822 - val_loss: 12.3066 - val_accuracy: 0.1000
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.9325 - accuracy: 0.3223
Epoch 39: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.9325 - accuracy: 0.3223 - val_loss: 2.4612 - val_accuracy: 0.2062
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.8898 - accuracy: 0.3311
Epoch 40: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.8898 - accuracy: 0.3311 - val_loss: 2.4262 - val_accuracy: 0.2250
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.8883 - accuracy: 0.3174
Epoch 41: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 103ms/step - loss: 1.8883 - accuracy: 0.3174 - val_loss: 3.8126 - val_accuracy: 0.2188
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.9383 - accuracy: 0.3018
Epoch 42: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 104ms/step - loss: 1.9383 - accuracy: 0.3018 - val_loss: 2.9020 - val_accuracy: 0.1875
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.9055 - accuracy: 0.3252
Epoch 43: val_accuracy did not improve from 0.28750
64/64 [==============================] - 7s 104ms/step - loss: 1.9055 - accuracy: 0.3252 - val_loss: 2.1219 - val_accuracy: 0.2812
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.9478 - accuracy: 0.3047
Epoch 44: val_accuracy improved from 0.28750 to 0.30625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 129ms/step - loss: 1.9478 - accuracy: 0.3047 - val_loss: 2.5076 - val_accuracy: 0.3063
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.9177 - accuracy: 0.3242
Epoch 45: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 105ms/step - loss: 1.9177 - accuracy: 0.3242 - val_loss: 3.6204 - val_accuracy: 0.2688
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.8720 - accuracy: 0.3223
Epoch 46: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 104ms/step - loss: 1.8720 - accuracy: 0.3223 - val_loss: 2.3644 - val_accuracy: 0.2750
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.8959 - accuracy: 0.3301
Epoch 47: val_accuracy improved from 0.30625 to 0.31875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 123ms/step - loss: 1.8959 - accuracy: 0.3301 - val_loss: 2.0881 - val_accuracy: 0.3187
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.8298 - accuracy: 0.3379
Epoch 48: val_accuracy improved from 0.31875 to 0.34375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 123ms/step - loss: 1.8298 - accuracy: 0.3379 - val_loss: 2.1810 - val_accuracy: 0.3438
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.8486 - accuracy: 0.3223
Epoch 49: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 104ms/step - loss: 1.8486 - accuracy: 0.3223 - val_loss: 2.0128 - val_accuracy: 0.3063
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.8809 - accuracy: 0.3213
Epoch 50: val_accuracy improved from 0.34375 to 0.36250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 120ms/step - loss: 1.8809 - accuracy: 0.3213 - val_loss: 2.3987 - val_accuracy: 0.3625
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.7984 - accuracy: 0.3574
Epoch 51: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 105ms/step - loss: 1.7984 - accuracy: 0.3574 - val_loss: 2.6725 - val_accuracy: 0.2375
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.8586 - accuracy: 0.3428
Epoch 52: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 104ms/step - loss: 1.8586 - accuracy: 0.3428 - val_loss: 2.2984 - val_accuracy: 0.3063
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.8844 - accuracy: 0.3340
Epoch 53: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 104ms/step - loss: 1.8844 - accuracy: 0.3340 - val_loss: 2.4855 - val_accuracy: 0.2438
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.8220 - accuracy: 0.3350
Epoch 54: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 106ms/step - loss: 1.8220 - accuracy: 0.3350 - val_loss: 1.8485 - val_accuracy: 0.3562
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.8297 - accuracy: 0.3379
Epoch 55: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 109ms/step - loss: 1.8297 - accuracy: 0.3379 - val_loss: 2.0261 - val_accuracy: 0.3125
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.8248 - accuracy: 0.3555
Epoch 56: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 109ms/step - loss: 1.8248 - accuracy: 0.3555 - val_loss: 2.0675 - val_accuracy: 0.3375
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.8080 - accuracy: 0.3574
Epoch 57: val_accuracy did not improve from 0.36250
64/64 [==============================] - 7s 109ms/step - loss: 1.8080 - accuracy: 0.3574 - val_loss: 3.1825 - val_accuracy: 0.2812
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.8447 - accuracy: 0.3643
Epoch 58: val_accuracy improved from 0.36250 to 0.39375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 9s 134ms/step - loss: 1.8447 - accuracy: 0.3643 - val_loss: 1.6772 - val_accuracy: 0.3938
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.8085 - accuracy: 0.3652
Epoch 59: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 113ms/step - loss: 1.8085 - accuracy: 0.3652 - val_loss: 4.8381 - val_accuracy: 0.2250
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.8360 - accuracy: 0.3389
Epoch 60: val_accuracy improved from 0.39375 to 0.41875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 129ms/step - loss: 1.8360 - accuracy: 0.3389 - val_loss: 1.7219 - val_accuracy: 0.4187
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.7985 - accuracy: 0.3457
Epoch 61: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.7985 - accuracy: 0.3457 - val_loss: 1.8486 - val_accuracy: 0.4062
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.8206 - accuracy: 0.3428
Epoch 62: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 112ms/step - loss: 1.8206 - accuracy: 0.3428 - val_loss: 2.7591 - val_accuracy: 0.3000
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.8268 - accuracy: 0.3564
Epoch 63: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 110ms/step - loss: 1.8268 - accuracy: 0.3564 - val_loss: 2.5044 - val_accuracy: 0.3250
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.8223 - accuracy: 0.3398
Epoch 64: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 117ms/step - loss: 1.8223 - accuracy: 0.3398 - val_loss: 2.3936 - val_accuracy: 0.3250
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.7765 - accuracy: 0.3623
Epoch 65: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 111ms/step - loss: 1.7765 - accuracy: 0.3623 - val_loss: 2.2138 - val_accuracy: 0.2937
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.7692 - accuracy: 0.3818
Epoch 66: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 111ms/step - loss: 1.7692 - accuracy: 0.3818 - val_loss: 1.8702 - val_accuracy: 0.3187
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.8092 - accuracy: 0.3721
Epoch 67: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 111ms/step - loss: 1.8092 - accuracy: 0.3721 - val_loss: 2.1350 - val_accuracy: 0.3063
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.7941 - accuracy: 0.3496
Epoch 68: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.7941 - accuracy: 0.3496 - val_loss: 1.9316 - val_accuracy: 0.2875
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.7438 - accuracy: 0.3887
Epoch 69: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.7438 - accuracy: 0.3887 - val_loss: 1.7607 - val_accuracy: 0.3375
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.7641 - accuracy: 0.3926
Epoch 70: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 115ms/step - loss: 1.7641 - accuracy: 0.3926 - val_loss: 1.7295 - val_accuracy: 0.4000
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.6805 - accuracy: 0.4121
Epoch 71: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6805 - accuracy: 0.4121 - val_loss: 3.7972 - val_accuracy: 0.2500
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.7743 - accuracy: 0.3965
Epoch 72: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.7743 - accuracy: 0.3965 - val_loss: 1.9582 - val_accuracy: 0.2875
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.7161 - accuracy: 0.4023
Epoch 73: val_accuracy did not improve from 0.41875
64/64 [==============================] - 8s 117ms/step - loss: 1.7161 - accuracy: 0.4023 - val_loss: 2.6636 - val_accuracy: 0.4125
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.7326 - accuracy: 0.3857
Epoch 74: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.7326 - accuracy: 0.3857 - val_loss: 1.9660 - val_accuracy: 0.3625
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.6958 - accuracy: 0.4102
Epoch 75: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 110ms/step - loss: 1.6958 - accuracy: 0.4102 - val_loss: 2.4611 - val_accuracy: 0.3875
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.7189 - accuracy: 0.4062
Epoch 76: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 116ms/step - loss: 1.7189 - accuracy: 0.4062 - val_loss: 1.9891 - val_accuracy: 0.3688
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.7217 - accuracy: 0.4023
Epoch 77: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 111ms/step - loss: 1.7217 - accuracy: 0.4023 - val_loss: 3.2608 - val_accuracy: 0.2250
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.6913 - accuracy: 0.4277
Epoch 78: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 111ms/step - loss: 1.6913 - accuracy: 0.4277 - val_loss: 5.0849 - val_accuracy: 0.2062
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.7674 - accuracy: 0.3711
Epoch 79: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 114ms/step - loss: 1.7674 - accuracy: 0.3711 - val_loss: 2.8523 - val_accuracy: 0.2500
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.6924 - accuracy: 0.4062
Epoch 80: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 112ms/step - loss: 1.6924 - accuracy: 0.4062 - val_loss: 3.0158 - val_accuracy: 0.3063
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.6796 - accuracy: 0.4141
Epoch 81: val_accuracy improved from 0.41875 to 0.45625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 128ms/step - loss: 1.6796 - accuracy: 0.4141 - val_loss: 1.5853 - val_accuracy: 0.4563
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.6614 - accuracy: 0.4111
Epoch 82: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 109ms/step - loss: 1.6614 - accuracy: 0.4111 - val_loss: 4.5780 - val_accuracy: 0.3375
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.6871 - accuracy: 0.4180
Epoch 83: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 109ms/step - loss: 1.6871 - accuracy: 0.4180 - val_loss: 4.4917 - val_accuracy: 0.2188
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.6745 - accuracy: 0.4043
Epoch 84: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 108ms/step - loss: 1.6745 - accuracy: 0.4043 - val_loss: 2.8306 - val_accuracy: 0.3125
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.6063 - accuracy: 0.4385
Epoch 85: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 108ms/step - loss: 1.6063 - accuracy: 0.4385 - val_loss: 1.8743 - val_accuracy: 0.3688
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.6923 - accuracy: 0.4346
Epoch 86: val_accuracy did not improve from 0.45625
64/64 [==============================] - 7s 109ms/step - loss: 1.6923 - accuracy: 0.4346 - val_loss: 2.2131 - val_accuracy: 0.3313
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.7203 - accuracy: 0.3945
Epoch 87: val_accuracy improved from 0.45625 to 0.51250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 129ms/step - loss: 1.7203 - accuracy: 0.3945 - val_loss: 1.5148 - val_accuracy: 0.5125
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.6009 - accuracy: 0.4385
Epoch 88: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 108ms/step - loss: 1.6009 - accuracy: 0.4385 - val_loss: 2.0619 - val_accuracy: 0.3562
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.6140 - accuracy: 0.4307
Epoch 89: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 108ms/step - loss: 1.6140 - accuracy: 0.4307 - val_loss: 2.7040 - val_accuracy: 0.3000
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.6442 - accuracy: 0.4258
Epoch 90: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 109ms/step - loss: 1.6442 - accuracy: 0.4258 - val_loss: 5.8102 - val_accuracy: 0.2000
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.5978 - accuracy: 0.4492
Epoch 91: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 110ms/step - loss: 1.5978 - accuracy: 0.4492 - val_loss: 1.4858 - val_accuracy: 0.4563
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.6437 - accuracy: 0.4404
Epoch 92: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 108ms/step - loss: 1.6437 - accuracy: 0.4404 - val_loss: 2.0178 - val_accuracy: 0.3438
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.5956 - accuracy: 0.4658
Epoch 93: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 110ms/step - loss: 1.5956 - accuracy: 0.4658 - val_loss: 1.8730 - val_accuracy: 0.3875
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.6105 - accuracy: 0.4346
Epoch 94: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 114ms/step - loss: 1.6105 - accuracy: 0.4346 - val_loss: 2.2021 - val_accuracy: 0.4000
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.6090 - accuracy: 0.4424
Epoch 95: val_accuracy did not improve from 0.51250
64/64 [==============================] - 7s 111ms/step - loss: 1.6090 - accuracy: 0.4424 - val_loss: 1.8643 - val_accuracy: 0.4000
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.5793 - accuracy: 0.4219
Epoch 96: val_accuracy improved from 0.51250 to 0.53750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_3.h5
64/64 [==============================] - 8s 128ms/step - loss: 1.5793 - accuracy: 0.4219 - val_loss: 1.3562 - val_accuracy: 0.5375
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.5527 - accuracy: 0.4668
Epoch 97: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.5527 - accuracy: 0.4668 - val_loss: 7.9603 - val_accuracy: 0.1562
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.5475 - accuracy: 0.4512
Epoch 98: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 109ms/step - loss: 1.5475 - accuracy: 0.4512 - val_loss: 2.3336 - val_accuracy: 0.3375
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.5688 - accuracy: 0.4502
Epoch 99: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 109ms/step - loss: 1.5688 - accuracy: 0.4502 - val_loss: 2.7253 - val_accuracy: 0.3313
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.4839 - accuracy: 0.4824
Epoch 100: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.4839 - accuracy: 0.4824 - val_loss: 5.2123 - val_accuracy: 0.2125
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.5696 - accuracy: 0.4463
Epoch 101: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.5696 - accuracy: 0.4463 - val_loss: 1.7663 - val_accuracy: 0.3750
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.4943 - accuracy: 0.4951
Epoch 102: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.4943 - accuracy: 0.4951 - val_loss: 6.1871 - val_accuracy: 0.2000
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.5635 - accuracy: 0.4590
Epoch 103: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.5635 - accuracy: 0.4590 - val_loss: 1.9740 - val_accuracy: 0.4062
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.4913 - accuracy: 0.5088
Epoch 104: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.4913 - accuracy: 0.5088 - val_loss: 7.5735 - val_accuracy: 0.1688
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.4838 - accuracy: 0.4707
Epoch 105: val_accuracy did not improve from 0.53750
64/64 [==============================] - 7s 108ms/step - loss: 1.4838 - accuracy: 0.4707 - val_loss: 8.0386 - val_accuracy: 0.1937


 ********* Training time: 1406.03125 s.


*****************
* Model Summary *
*****************
Model: "resnet50"
__________________________________________________________________________________________________
 Layer (type)                   Output Shape         Param #     Connected to                     
==================================================================================================
 input_3 (InputLayer)           [(None, 224, 224, 3  0           []                               
                                )]                                                                
                                                                                                  
 conv1_pad (ZeroPadding2D)      (None, 230, 230, 3)  0           ['input_3[0][0]']                
                                                                                                  
 conv1_conv (Conv2D)            (None, 112, 112, 64  9472        ['conv1_pad[0][0]']              
                                )                                                                 
                                                                                                  
 conv1_bn (BatchNormalization)  (None, 112, 112, 64  256         ['conv1_conv[0][0]']             
                                )                                                                 
                                                                                                  
 conv1_relu (Activation)        (None, 112, 112, 64  0           ['conv1_bn[0][0]']               
                                )                                                                 
                                                                                                  
 pool1_pad (ZeroPadding2D)      (None, 114, 114, 64  0           ['conv1_relu[0][0]']             
                                )                                                                 
                                                                                                  
 pool1_pool (MaxPooling2D)      (None, 56, 56, 64)   0           ['pool1_pad[0][0]']              
                                                                                                  
 conv2_block1_1_conv (Conv2D)   (None, 56, 56, 64)   4160        ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block1_1_relu[0][0]']    
                                                                                                  
 conv2_block1_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_0_conv (Conv2D)   (None, 56, 56, 256)  16640       ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block1_2_relu[0][0]']    
                                                                                                  
 conv2_block1_0_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_0_bn[0][0]',      
                                                                  'conv2_block1_3_bn[0][0]']      
                                                                                                  
 conv2_block1_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block1_add[0][0]']       
                                                                                                  
 conv2_block2_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block1_out[0][0]']       
                                                                                                  
 conv2_block2_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block2_1_relu[0][0]']    
                                                                                                  
 conv2_block2_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block2_2_relu[0][0]']    
                                                                                                  
 conv2_block2_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_out[0][0]',       
                                                                  'conv2_block2_3_bn[0][0]']      
                                                                                                  
 conv2_block2_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block2_add[0][0]']       
                                                                                                  
 conv2_block3_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block2_out[0][0]']       
                                                                                                  
 conv2_block3_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block3_1_relu[0][0]']    
                                                                                                  
 conv2_block3_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block3_2_relu[0][0]']    
                                                                                                  
 conv2_block3_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_add (Add)         (None, 56, 56, 256)  0           ['conv2_block2_out[0][0]',       
                                                                  'conv2_block3_3_bn[0][0]']      
                                                                                                  
 conv2_block3_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block3_add[0][0]']       
                                                                                                  
 conv3_block1_1_conv (Conv2D)   (None, 28, 28, 128)  32896       ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block1_1_relu[0][0]']    
                                                                                                  
 conv3_block1_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_0_conv (Conv2D)   (None, 28, 28, 512)  131584      ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block1_2_relu[0][0]']    
                                                                                                  
 conv3_block1_0_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_0_bn[0][0]',      
                                                                  'conv3_block1_3_bn[0][0]']      
                                                                                                  
 conv3_block1_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block1_add[0][0]']       
                                                                                                  
 conv3_block2_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block1_out[0][0]']       
                                                                                                  
 conv3_block2_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block2_1_relu[0][0]']    
                                                                                                  
 conv3_block2_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block2_2_relu[0][0]']    
                                                                                                  
 conv3_block2_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_out[0][0]',       
                                                                  'conv3_block2_3_bn[0][0]']      
                                                                                                  
 conv3_block2_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block2_add[0][0]']       
                                                                                                  
 conv3_block3_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block2_out[0][0]']       
                                                                                                  
 conv3_block3_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block3_1_relu[0][0]']    
                                                                                                  
 conv3_block3_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block3_2_relu[0][0]']    
                                                                                                  
 conv3_block3_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_add (Add)         (None, 28, 28, 512)  0           ['conv3_block2_out[0][0]',       
                                                                  'conv3_block3_3_bn[0][0]']      
                                                                                                  
 conv3_block3_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block3_add[0][0]']       
                                                                                                  
 conv3_block4_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block3_out[0][0]']       
                                                                                                  
 conv3_block4_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block4_1_relu[0][0]']    
                                                                                                  
 conv3_block4_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block4_2_relu[0][0]']    
                                                                                                  
 conv3_block4_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block4_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_add (Add)         (None, 28, 28, 512)  0           ['conv3_block3_out[0][0]',       
                                                                  'conv3_block4_3_bn[0][0]']      
                                                                                                  
 conv3_block4_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block4_add[0][0]']       
                                                                                                  
 conv4_block1_1_conv (Conv2D)   (None, 14, 14, 256)  131328      ['conv3_block4_out[0][0]']       
                                                                                                  
 conv4_block1_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block1_1_relu[0][0]']    
                                                                                                  
 conv4_block1_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_0_conv (Conv2D)   (None, 14, 14, 1024  525312      ['conv3_block4_out[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block1_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block1_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block1_0_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_0_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_0_bn[0][0]',      
                                )                                 'conv4_block1_3_bn[0][0]']      
                                                                                                  
 conv4_block1_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block1_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block2_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block1_out[0][0]']       
                                                                                                  
 conv4_block2_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block2_1_relu[0][0]']    
                                                                                                  
 conv4_block2_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block2_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block2_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block2_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block2_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_out[0][0]',       
                                )                                 'conv4_block2_3_bn[0][0]']      
                                                                                                  
 conv4_block2_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block2_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block3_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block2_out[0][0]']       
                                                                                                  
 conv4_block3_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block3_1_relu[0][0]']    
                                                                                                  
 conv4_block3_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block3_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block3_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block3_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block3_add (Add)         (None, 14, 14, 1024  0           ['conv4_block2_out[0][0]',       
                                )                                 'conv4_block3_3_bn[0][0]']      
                                                                                                  
 conv4_block3_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block3_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block4_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block3_out[0][0]']       
                                                                                                  
 conv4_block4_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block4_1_relu[0][0]']    
                                                                                                  
 conv4_block4_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block4_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block4_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block4_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block4_add (Add)         (None, 14, 14, 1024  0           ['conv4_block3_out[0][0]',       
                                )                                 'conv4_block4_3_bn[0][0]']      
                                                                                                  
 conv4_block4_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block4_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block5_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block4_out[0][0]']       
                                                                                                  
 conv4_block5_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block5_1_relu[0][0]']    
                                                                                                  
 conv4_block5_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block5_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block5_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block5_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block5_add (Add)         (None, 14, 14, 1024  0           ['conv4_block4_out[0][0]',       
                                )                                 'conv4_block5_3_bn[0][0]']      
                                                                                                  
 conv4_block5_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block5_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block6_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block5_out[0][0]']       
                                                                                                  
 conv4_block6_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block6_1_relu[0][0]']    
                                                                                                  
 conv4_block6_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block6_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block6_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block6_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block6_add (Add)         (None, 14, 14, 1024  0           ['conv4_block5_out[0][0]',       
                                )                                 'conv4_block6_3_bn[0][0]']      
                                                                                                  
 conv4_block6_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block6_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv5_block1_1_conv (Conv2D)   (None, 7, 7, 512)    524800      ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block1_1_relu[0][0]']    
                                                                                                  
 conv5_block1_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_0_conv (Conv2D)   (None, 7, 7, 2048)   2099200     ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block1_2_relu[0][0]']    
                                                                                                  
 conv5_block1_0_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_0_bn[0][0]',      
                                                                  'conv5_block1_3_bn[0][0]']      
                                                                                                  
 conv5_block1_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block1_add[0][0]']       
                                                                                                  
 conv5_block2_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block1_out[0][0]']       
                                                                                                  
 conv5_block2_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block2_1_relu[0][0]']    
                                                                                                  
 conv5_block2_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block2_2_relu[0][0]']    
                                                                                                  
 conv5_block2_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_out[0][0]',       
                                                                  'conv5_block2_3_bn[0][0]']      
                                                                                                  
 conv5_block2_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block2_add[0][0]']       
                                                                                                  
 conv5_block3_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block2_out[0][0]']       
                                                                                                  
 conv5_block3_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block3_1_relu[0][0]']    
                                                                                                  
 conv5_block3_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block3_2_relu[0][0]']    
                                                                                                  
 conv5_block3_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block2_out[0][0]',       
                                                                  'conv5_block3_3_bn[0][0]']      
                                                                                                  
 conv5_block3_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block3_add[0][0]']       
                                                                                                  
 avg_pool (GlobalAveragePooling  (None, 2048)        0           ['conv5_block3_out[0][0]']       
 2D)                                                                                              
                                                                                                  
 predictions (Dense)            (None, 12)           24588       ['avg_pool[0][0]']               
                                                                                                  
==================================================================================================
Total params: 23,612,300
Trainable params: 23,559,180
Non-trainable params: 53,120
__________________________________________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 27ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.3663    0.3964    0.3808       280
anarrhichomenum     0.1085    1.0000    0.1958       319
   brevantherum     0.0000    0.0000    0.0000       267
      dulcamara     0.2900    0.0921    0.1398       315
   herposolanum     0.5946    0.0733    0.1306       300
     holophylla     0.4286    0.0537    0.0955       335
     lasiocarpa     0.9153    0.1748    0.2935       309
      melongena     0.5000    0.0035    0.0070       285
    micracantha     0.2131    0.0442    0.0732       294
         petota     0.1143    0.0133    0.0238       301
        solanum     0.2500    0.0033    0.0066       299
          torva     0.4118    0.0236    0.0447       296

       accuracy                         0.1608      3600
      macro avg     0.3494    0.1565    0.1159      3600
   weighted avg     0.3526    0.1608    0.1169      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

****************************
* Started at 4290.03125... *
****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 3.9890 - accuracy: 0.0918
Epoch 1: val_accuracy improved from -inf to 0.08750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 12s 125ms/step - loss: 3.9890 - accuracy: 0.0918 - val_loss: 56.2198 - val_accuracy: 0.0875
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 2.8317 - accuracy: 0.1094
Epoch 2: val_accuracy did not improve from 0.08750
64/64 [==============================] - 7s 108ms/step - loss: 2.8317 - accuracy: 0.1094 - val_loss: 1518.9069 - val_accuracy: 0.0750
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 2.6752 - accuracy: 0.1299
Epoch 3: val_accuracy improved from 0.08750 to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.6752 - accuracy: 0.1299 - val_loss: 11.1196 - val_accuracy: 0.1063
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 2.5130 - accuracy: 0.1328
Epoch 4: val_accuracy did not improve from 0.10625
64/64 [==============================] - 7s 109ms/step - loss: 2.5130 - accuracy: 0.1328 - val_loss: 213.6284 - val_accuracy: 0.0812
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 2.5858 - accuracy: 0.1230
Epoch 5: val_accuracy did not improve from 0.10625
64/64 [==============================] - 7s 109ms/step - loss: 2.5858 - accuracy: 0.1230 - val_loss: 42.7997 - val_accuracy: 0.1000
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.4902 - accuracy: 0.1260
Epoch 6: val_accuracy did not improve from 0.10625
64/64 [==============================] - 7s 108ms/step - loss: 2.4902 - accuracy: 0.1260 - val_loss: 2.9617 - val_accuracy: 0.1000
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.4348 - accuracy: 0.1582
Epoch 7: val_accuracy improved from 0.10625 to 0.11875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 131ms/step - loss: 2.4348 - accuracy: 0.1582 - val_loss: 25.0416 - val_accuracy: 0.1187
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.4074 - accuracy: 0.1406
Epoch 8: val_accuracy did not improve from 0.11875
64/64 [==============================] - 7s 108ms/step - loss: 2.4074 - accuracy: 0.1406 - val_loss: 5.2432 - val_accuracy: 0.1063
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.3631 - accuracy: 0.1680
Epoch 9: val_accuracy did not improve from 0.11875
64/64 [==============================] - 7s 108ms/step - loss: 2.3631 - accuracy: 0.1680 - val_loss: 26.4269 - val_accuracy: 0.0437
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.2893 - accuracy: 0.1855
Epoch 10: val_accuracy did not improve from 0.11875
64/64 [==============================] - 7s 108ms/step - loss: 2.2893 - accuracy: 0.1855 - val_loss: 14.9559 - val_accuracy: 0.0562
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.2328 - accuracy: 0.1807
Epoch 11: val_accuracy improved from 0.11875 to 0.13125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 129ms/step - loss: 2.2328 - accuracy: 0.1807 - val_loss: 3.4483 - val_accuracy: 0.1312
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.2464 - accuracy: 0.1816
Epoch 12: val_accuracy improved from 0.13125 to 0.17500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 9s 133ms/step - loss: 2.2464 - accuracy: 0.1816 - val_loss: 5.0422 - val_accuracy: 0.1750
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.2326 - accuracy: 0.1963
Epoch 13: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 109ms/step - loss: 2.2326 - accuracy: 0.1963 - val_loss: 20.9180 - val_accuracy: 0.1375
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.1849 - accuracy: 0.2012
Epoch 14: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 109ms/step - loss: 2.1849 - accuracy: 0.2012 - val_loss: 3.1571 - val_accuracy: 0.1437
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.2083 - accuracy: 0.2041
Epoch 15: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 108ms/step - loss: 2.2083 - accuracy: 0.2041 - val_loss: 4.4903 - val_accuracy: 0.1750
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.1849 - accuracy: 0.2109
Epoch 16: val_accuracy improved from 0.17500 to 0.18750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 9s 136ms/step - loss: 2.1849 - accuracy: 0.2109 - val_loss: 2.4158 - val_accuracy: 0.1875
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.1788 - accuracy: 0.2070
Epoch 17: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 108ms/step - loss: 2.1788 - accuracy: 0.2070 - val_loss: 2.6501 - val_accuracy: 0.1813
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.1325 - accuracy: 0.2197
Epoch 18: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 108ms/step - loss: 2.1325 - accuracy: 0.2197 - val_loss: 2.4996 - val_accuracy: 0.1625
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.1183 - accuracy: 0.2188
Epoch 19: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 108ms/step - loss: 2.1183 - accuracy: 0.2188 - val_loss: 4.5839 - val_accuracy: 0.0812
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.1559 - accuracy: 0.2207
Epoch 20: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 109ms/step - loss: 2.1559 - accuracy: 0.2207 - val_loss: 2.2319 - val_accuracy: 0.1875
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.0922 - accuracy: 0.2109
Epoch 21: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 115ms/step - loss: 2.0922 - accuracy: 0.2109 - val_loss: 2.7259 - val_accuracy: 0.1000
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.1204 - accuracy: 0.2021
Epoch 22: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 110ms/step - loss: 2.1204 - accuracy: 0.2021 - val_loss: 2.2841 - val_accuracy: 0.1813
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.0776 - accuracy: 0.2461
Epoch 23: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 109ms/step - loss: 2.0776 - accuracy: 0.2461 - val_loss: 4.1128 - val_accuracy: 0.1375
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.1027 - accuracy: 0.2314
Epoch 24: val_accuracy improved from 0.18750 to 0.20625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 9s 136ms/step - loss: 2.1027 - accuracy: 0.2314 - val_loss: 2.9314 - val_accuracy: 0.2062
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.0084 - accuracy: 0.2773
Epoch 25: val_accuracy did not improve from 0.20625
64/64 [==============================] - 7s 108ms/step - loss: 2.0084 - accuracy: 0.2773 - val_loss: 3.2564 - val_accuracy: 0.1875
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.0123 - accuracy: 0.2725
Epoch 26: val_accuracy did not improve from 0.20625
64/64 [==============================] - 7s 108ms/step - loss: 2.0123 - accuracy: 0.2725 - val_loss: 4.4920 - val_accuracy: 0.1813
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.0375 - accuracy: 0.2666
Epoch 27: val_accuracy did not improve from 0.20625
64/64 [==============================] - 7s 110ms/step - loss: 2.0375 - accuracy: 0.2666 - val_loss: 6.9921 - val_accuracy: 0.1250
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.0064 - accuracy: 0.2568
Epoch 28: val_accuracy improved from 0.20625 to 0.25625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 125ms/step - loss: 2.0064 - accuracy: 0.2568 - val_loss: 2.3767 - val_accuracy: 0.2562
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 1.9949 - accuracy: 0.2686
Epoch 29: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 108ms/step - loss: 1.9949 - accuracy: 0.2686 - val_loss: 4.1448 - val_accuracy: 0.1688
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.0360 - accuracy: 0.2568
Epoch 30: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 108ms/step - loss: 2.0360 - accuracy: 0.2568 - val_loss: 4.0018 - val_accuracy: 0.1750
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 1.9382 - accuracy: 0.3145
Epoch 31: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 109ms/step - loss: 1.9382 - accuracy: 0.3145 - val_loss: 2.1866 - val_accuracy: 0.2375
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.0084 - accuracy: 0.2666
Epoch 32: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 109ms/step - loss: 2.0084 - accuracy: 0.2666 - val_loss: 2.1756 - val_accuracy: 0.2313
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 1.9630 - accuracy: 0.3047
Epoch 33: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 115ms/step - loss: 1.9630 - accuracy: 0.3047 - val_loss: 2.9268 - val_accuracy: 0.2125
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 2.0253 - accuracy: 0.2500
Epoch 34: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 112ms/step - loss: 2.0253 - accuracy: 0.2500 - val_loss: 3.5004 - val_accuracy: 0.1750
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 1.9954 - accuracy: 0.2725
Epoch 35: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 109ms/step - loss: 1.9954 - accuracy: 0.2725 - val_loss: 4.1170 - val_accuracy: 0.1937
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 1.9349 - accuracy: 0.3379
Epoch 36: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 110ms/step - loss: 1.9349 - accuracy: 0.3379 - val_loss: 3.4796 - val_accuracy: 0.1937
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 2.0085 - accuracy: 0.2773
Epoch 37: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 108ms/step - loss: 2.0085 - accuracy: 0.2773 - val_loss: 3.4351 - val_accuracy: 0.2375
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 1.9349 - accuracy: 0.2979
Epoch 38: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 108ms/step - loss: 1.9349 - accuracy: 0.2979 - val_loss: 2.8473 - val_accuracy: 0.2125
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 1.9304 - accuracy: 0.3057
Epoch 39: val_accuracy did not improve from 0.25625
64/64 [==============================] - 7s 108ms/step - loss: 1.9304 - accuracy: 0.3057 - val_loss: 4.4854 - val_accuracy: 0.1813
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 1.9390 - accuracy: 0.2979
Epoch 40: val_accuracy improved from 0.25625 to 0.35000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 128ms/step - loss: 1.9390 - accuracy: 0.2979 - val_loss: 1.9167 - val_accuracy: 0.3500
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 1.9106 - accuracy: 0.3076
Epoch 41: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.9106 - accuracy: 0.3076 - val_loss: 2.2609 - val_accuracy: 0.2250
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 1.9191 - accuracy: 0.3174
Epoch 42: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 107ms/step - loss: 1.9191 - accuracy: 0.3174 - val_loss: 2.8451 - val_accuracy: 0.2062
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 1.9054 - accuracy: 0.2998
Epoch 43: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.9054 - accuracy: 0.2998 - val_loss: 4.2945 - val_accuracy: 0.1875
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 1.8857 - accuracy: 0.3125
Epoch 44: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.8857 - accuracy: 0.3125 - val_loss: 1.8480 - val_accuracy: 0.3438
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 1.8791 - accuracy: 0.3066
Epoch 45: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.8791 - accuracy: 0.3066 - val_loss: 2.2119 - val_accuracy: 0.2875
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 1.9720 - accuracy: 0.2930
Epoch 46: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 107ms/step - loss: 1.9720 - accuracy: 0.2930 - val_loss: 5.0282 - val_accuracy: 0.2188
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 1.9373 - accuracy: 0.3057
Epoch 47: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 107ms/step - loss: 1.9373 - accuracy: 0.3057 - val_loss: 3.5290 - val_accuracy: 0.2125
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 1.8992 - accuracy: 0.3232
Epoch 48: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 107ms/step - loss: 1.8992 - accuracy: 0.3232 - val_loss: 3.0855 - val_accuracy: 0.2438
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 1.9142 - accuracy: 0.3184
Epoch 49: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.9142 - accuracy: 0.3184 - val_loss: 2.0345 - val_accuracy: 0.3000
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 1.9081 - accuracy: 0.3320
Epoch 50: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 1.9081 - accuracy: 0.3320 - val_loss: 2.3816 - val_accuracy: 0.2625
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 1.8490 - accuracy: 0.3359
Epoch 51: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.8490 - accuracy: 0.3359 - val_loss: 2.5539 - val_accuracy: 0.3000
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.8251 - accuracy: 0.3447
Epoch 52: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.8251 - accuracy: 0.3447 - val_loss: 2.0784 - val_accuracy: 0.3438
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 1.8323 - accuracy: 0.3535
Epoch 53: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 114ms/step - loss: 1.8323 - accuracy: 0.3535 - val_loss: 3.0686 - val_accuracy: 0.2375
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 1.8940 - accuracy: 0.3359
Epoch 54: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.8940 - accuracy: 0.3359 - val_loss: 2.0675 - val_accuracy: 0.2937
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 1.8594 - accuracy: 0.3516
Epoch 55: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 1.8594 - accuracy: 0.3516 - val_loss: 3.6510 - val_accuracy: 0.2125
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 1.9154 - accuracy: 0.3086
Epoch 56: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 114ms/step - loss: 1.9154 - accuracy: 0.3086 - val_loss: 1.8295 - val_accuracy: 0.3500
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.8954 - accuracy: 0.3232
Epoch 57: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 1.8954 - accuracy: 0.3232 - val_loss: 2.3479 - val_accuracy: 0.2000
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 1.7845 - accuracy: 0.3506
Epoch 58: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 1.7845 - accuracy: 0.3506 - val_loss: 2.1938 - val_accuracy: 0.2750
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 1.8423 - accuracy: 0.3359
Epoch 59: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 115ms/step - loss: 1.8423 - accuracy: 0.3359 - val_loss: 1.9672 - val_accuracy: 0.2438
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.8221 - accuracy: 0.3525
Epoch 60: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 115ms/step - loss: 1.8221 - accuracy: 0.3525 - val_loss: 2.4610 - val_accuracy: 0.2125
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.7808 - accuracy: 0.3730
Epoch 61: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.7808 - accuracy: 0.3730 - val_loss: 2.1835 - val_accuracy: 0.2937
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 1.8701 - accuracy: 0.3467
Epoch 62: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 1.8701 - accuracy: 0.3467 - val_loss: 2.3839 - val_accuracy: 0.3125
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.8164 - accuracy: 0.3516
Epoch 63: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 114ms/step - loss: 1.8164 - accuracy: 0.3516 - val_loss: 4.6246 - val_accuracy: 0.1625
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 1.7910 - accuracy: 0.3701
Epoch 64: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 1.7910 - accuracy: 0.3701 - val_loss: 2.0328 - val_accuracy: 0.3000
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.7716 - accuracy: 0.3770
Epoch 65: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7716 - accuracy: 0.3770 - val_loss: 3.1629 - val_accuracy: 0.2125
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.8761 - accuracy: 0.3359
Epoch 66: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.8761 - accuracy: 0.3359 - val_loss: 2.4284 - val_accuracy: 0.2937
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.8640 - accuracy: 0.3389
Epoch 67: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.8640 - accuracy: 0.3389 - val_loss: 2.4586 - val_accuracy: 0.2562
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 1.7944 - accuracy: 0.3564
Epoch 68: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.7944 - accuracy: 0.3564 - val_loss: 2.0443 - val_accuracy: 0.3438
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.8779 - accuracy: 0.3398
Epoch 69: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.8779 - accuracy: 0.3398 - val_loss: 8.6222 - val_accuracy: 0.1312
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.7559 - accuracy: 0.3799
Epoch 70: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 108ms/step - loss: 1.7559 - accuracy: 0.3799 - val_loss: 4.6273 - val_accuracy: 0.2000
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.7605 - accuracy: 0.3945
Epoch 71: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 1.7605 - accuracy: 0.3945 - val_loss: 2.1680 - val_accuracy: 0.3500
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.8384 - accuracy: 0.3311
Epoch 72: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 1.8384 - accuracy: 0.3311 - val_loss: 1.8988 - val_accuracy: 0.3125
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.7392 - accuracy: 0.3887
Epoch 73: val_accuracy improved from 0.35000 to 0.39375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 8s 127ms/step - loss: 1.7392 - accuracy: 0.3887 - val_loss: 1.7783 - val_accuracy: 0.3938
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.8174 - accuracy: 0.3643
Epoch 74: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 110ms/step - loss: 1.8174 - accuracy: 0.3643 - val_loss: 2.1298 - val_accuracy: 0.2812
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.7179 - accuracy: 0.4102
Epoch 75: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 115ms/step - loss: 1.7179 - accuracy: 0.4102 - val_loss: 2.5623 - val_accuracy: 0.2812
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.7200 - accuracy: 0.4014
Epoch 76: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 113ms/step - loss: 1.7200 - accuracy: 0.4014 - val_loss: 1.9940 - val_accuracy: 0.3313
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.7650 - accuracy: 0.3691
Epoch 77: val_accuracy did not improve from 0.39375
64/64 [==============================] - 8s 117ms/step - loss: 1.7650 - accuracy: 0.3691 - val_loss: 1.8903 - val_accuracy: 0.3187
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.7151 - accuracy: 0.3857
Epoch 78: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 113ms/step - loss: 1.7151 - accuracy: 0.3857 - val_loss: 2.1607 - val_accuracy: 0.2625
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.7118 - accuracy: 0.3896
Epoch 79: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 113ms/step - loss: 1.7118 - accuracy: 0.3896 - val_loss: 2.5579 - val_accuracy: 0.3125
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.7500 - accuracy: 0.3662
Epoch 80: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 112ms/step - loss: 1.7500 - accuracy: 0.3662 - val_loss: 1.8617 - val_accuracy: 0.3438
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.7402 - accuracy: 0.3838
Epoch 81: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 115ms/step - loss: 1.7402 - accuracy: 0.3838 - val_loss: 1.9147 - val_accuracy: 0.3063
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.7417 - accuracy: 0.3633
Epoch 82: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 110ms/step - loss: 1.7417 - accuracy: 0.3633 - val_loss: 2.1663 - val_accuracy: 0.3812
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.7048 - accuracy: 0.3945
Epoch 83: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 108ms/step - loss: 1.7048 - accuracy: 0.3945 - val_loss: 1.9891 - val_accuracy: 0.3313
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.7710 - accuracy: 0.3916
Epoch 84: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 109ms/step - loss: 1.7710 - accuracy: 0.3916 - val_loss: 2.0776 - val_accuracy: 0.3063
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.7036 - accuracy: 0.4033
Epoch 85: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 109ms/step - loss: 1.7036 - accuracy: 0.4033 - val_loss: 2.3543 - val_accuracy: 0.2562
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.6645 - accuracy: 0.4033
Epoch 86: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 109ms/step - loss: 1.6645 - accuracy: 0.4033 - val_loss: 3.6196 - val_accuracy: 0.2250
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.6858 - accuracy: 0.4131
Epoch 87: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 109ms/step - loss: 1.6858 - accuracy: 0.4131 - val_loss: 2.5743 - val_accuracy: 0.2937
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.6449 - accuracy: 0.4346
Epoch 88: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 110ms/step - loss: 1.6449 - accuracy: 0.4346 - val_loss: 2.9760 - val_accuracy: 0.2375
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.7367 - accuracy: 0.3916
Epoch 89: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 110ms/step - loss: 1.7367 - accuracy: 0.3916 - val_loss: 2.2156 - val_accuracy: 0.3438
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.6443 - accuracy: 0.4326
Epoch 90: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 110ms/step - loss: 1.6443 - accuracy: 0.4326 - val_loss: 1.9783 - val_accuracy: 0.3125
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.6505 - accuracy: 0.4092
Epoch 91: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 112ms/step - loss: 1.6505 - accuracy: 0.4092 - val_loss: 1.8269 - val_accuracy: 0.3688
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.6804 - accuracy: 0.4082
Epoch 92: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 111ms/step - loss: 1.6804 - accuracy: 0.4082 - val_loss: 1.8600 - val_accuracy: 0.3375
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.6439 - accuracy: 0.4297
Epoch 93: val_accuracy did not improve from 0.39375
64/64 [==============================] - 7s 115ms/step - loss: 1.6439 - accuracy: 0.4297 - val_loss: 2.2290 - val_accuracy: 0.2562
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.6275 - accuracy: 0.4365
Epoch 94: val_accuracy improved from 0.39375 to 0.41875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_20230216150341_4.h5
64/64 [==============================] - 9s 140ms/step - loss: 1.6275 - accuracy: 0.4365 - val_loss: 1.7236 - val_accuracy: 0.4187
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.6243 - accuracy: 0.4277
Epoch 95: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6243 - accuracy: 0.4277 - val_loss: 1.6285 - val_accuracy: 0.4187
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.6698 - accuracy: 0.4043
Epoch 96: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6698 - accuracy: 0.4043 - val_loss: 2.5748 - val_accuracy: 0.3250
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.6297 - accuracy: 0.4346
Epoch 97: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6297 - accuracy: 0.4346 - val_loss: 2.7248 - val_accuracy: 0.2875
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.6378 - accuracy: 0.4336
Epoch 98: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6378 - accuracy: 0.4336 - val_loss: 2.3096 - val_accuracy: 0.2937
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.5916 - accuracy: 0.4385
Epoch 99: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 108ms/step - loss: 1.5916 - accuracy: 0.4385 - val_loss: 2.4441 - val_accuracy: 0.2750
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.6416 - accuracy: 0.4395
Epoch 100: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 108ms/step - loss: 1.6416 - accuracy: 0.4395 - val_loss: 2.0823 - val_accuracy: 0.3250
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.6305 - accuracy: 0.4199
Epoch 101: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6305 - accuracy: 0.4199 - val_loss: 2.3706 - val_accuracy: 0.2812
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.6028 - accuracy: 0.4512
Epoch 102: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 108ms/step - loss: 1.6028 - accuracy: 0.4512 - val_loss: 2.0280 - val_accuracy: 0.3750
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.6113 - accuracy: 0.4395
Epoch 103: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.6113 - accuracy: 0.4395 - val_loss: 1.7363 - val_accuracy: 0.3688
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.5499 - accuracy: 0.4580
Epoch 104: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 108ms/step - loss: 1.5499 - accuracy: 0.4580 - val_loss: 2.1209 - val_accuracy: 0.3375
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.5882 - accuracy: 0.4580
Epoch 105: val_accuracy did not improve from 0.41875
64/64 [==============================] - 7s 109ms/step - loss: 1.5882 - accuracy: 0.4580 - val_loss: 1.9409 - val_accuracy: 0.4187


 ********* Training time: 1407.90625 s.


*****************
* Model Summary *
*****************
Model: "resnet50"
__________________________________________________________________________________________________
 Layer (type)                   Output Shape         Param #     Connected to                     
==================================================================================================
 input_4 (InputLayer)           [(None, 224, 224, 3  0           []                               
                                )]                                                                
                                                                                                  
 conv1_pad (ZeroPadding2D)      (None, 230, 230, 3)  0           ['input_4[0][0]']                
                                                                                                  
 conv1_conv (Conv2D)            (None, 112, 112, 64  9472        ['conv1_pad[0][0]']              
                                )                                                                 
                                                                                                  
 conv1_bn (BatchNormalization)  (None, 112, 112, 64  256         ['conv1_conv[0][0]']             
                                )                                                                 
                                                                                                  
 conv1_relu (Activation)        (None, 112, 112, 64  0           ['conv1_bn[0][0]']               
                                )                                                                 
                                                                                                  
 pool1_pad (ZeroPadding2D)      (None, 114, 114, 64  0           ['conv1_relu[0][0]']             
                                )                                                                 
                                                                                                  
 pool1_pool (MaxPooling2D)      (None, 56, 56, 64)   0           ['pool1_pad[0][0]']              
                                                                                                  
 conv2_block1_1_conv (Conv2D)   (None, 56, 56, 64)   4160        ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block1_1_relu[0][0]']    
                                                                                                  
 conv2_block1_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block1_0_conv (Conv2D)   (None, 56, 56, 256)  16640       ['pool1_pool[0][0]']             
                                                                                                  
 conv2_block1_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block1_2_relu[0][0]']    
                                                                                                  
 conv2_block1_0_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block1_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_0_bn[0][0]',      
                                                                  'conv2_block1_3_bn[0][0]']      
                                                                                                  
 conv2_block1_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block1_add[0][0]']       
                                                                                                  
 conv2_block2_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block1_out[0][0]']       
                                                                                                  
 conv2_block2_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block2_1_relu[0][0]']    
                                                                                                  
 conv2_block2_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block2_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block2_2_relu[0][0]']    
                                                                                                  
 conv2_block2_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block2_add (Add)         (None, 56, 56, 256)  0           ['conv2_block1_out[0][0]',       
                                                                  'conv2_block2_3_bn[0][0]']      
                                                                                                  
 conv2_block2_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block2_add[0][0]']       
                                                                                                  
 conv2_block3_1_conv (Conv2D)   (None, 56, 56, 64)   16448       ['conv2_block2_out[0][0]']       
                                                                                                  
 conv2_block3_1_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_1_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_2_conv (Conv2D)   (None, 56, 56, 64)   36928       ['conv2_block3_1_relu[0][0]']    
                                                                                                  
 conv2_block3_2_bn (BatchNormal  (None, 56, 56, 64)  256         ['conv2_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_2_relu (Activatio  (None, 56, 56, 64)  0           ['conv2_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv2_block3_3_conv (Conv2D)   (None, 56, 56, 256)  16640       ['conv2_block3_2_relu[0][0]']    
                                                                                                  
 conv2_block3_3_bn (BatchNormal  (None, 56, 56, 256)  1024       ['conv2_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv2_block3_add (Add)         (None, 56, 56, 256)  0           ['conv2_block2_out[0][0]',       
                                                                  'conv2_block3_3_bn[0][0]']      
                                                                                                  
 conv2_block3_out (Activation)  (None, 56, 56, 256)  0           ['conv2_block3_add[0][0]']       
                                                                                                  
 conv3_block1_1_conv (Conv2D)   (None, 28, 28, 128)  32896       ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block1_1_relu[0][0]']    
                                                                                                  
 conv3_block1_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block1_0_conv (Conv2D)   (None, 28, 28, 512)  131584      ['conv2_block3_out[0][0]']       
                                                                                                  
 conv3_block1_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block1_2_relu[0][0]']    
                                                                                                  
 conv3_block1_0_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block1_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_0_bn[0][0]',      
                                                                  'conv3_block1_3_bn[0][0]']      
                                                                                                  
 conv3_block1_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block1_add[0][0]']       
                                                                                                  
 conv3_block2_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block1_out[0][0]']       
                                                                                                  
 conv3_block2_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block2_1_relu[0][0]']    
                                                                                                  
 conv3_block2_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block2_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block2_2_relu[0][0]']    
                                                                                                  
 conv3_block2_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block2_add (Add)         (None, 28, 28, 512)  0           ['conv3_block1_out[0][0]',       
                                                                  'conv3_block2_3_bn[0][0]']      
                                                                                                  
 conv3_block2_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block2_add[0][0]']       
                                                                                                  
 conv3_block3_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block2_out[0][0]']       
                                                                                                  
 conv3_block3_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block3_1_relu[0][0]']    
                                                                                                  
 conv3_block3_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block3_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block3_2_relu[0][0]']    
                                                                                                  
 conv3_block3_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block3_add (Add)         (None, 28, 28, 512)  0           ['conv3_block2_out[0][0]',       
                                                                  'conv3_block3_3_bn[0][0]']      
                                                                                                  
 conv3_block3_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block3_add[0][0]']       
                                                                                                  
 conv3_block4_1_conv (Conv2D)   (None, 28, 28, 128)  65664       ['conv3_block3_out[0][0]']       
                                                                                                  
 conv3_block4_1_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_1_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_2_conv (Conv2D)   (None, 28, 28, 128)  147584      ['conv3_block4_1_relu[0][0]']    
                                                                                                  
 conv3_block4_2_bn (BatchNormal  (None, 28, 28, 128)  512        ['conv3_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_2_relu (Activatio  (None, 28, 28, 128)  0          ['conv3_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv3_block4_3_conv (Conv2D)   (None, 28, 28, 512)  66048       ['conv3_block4_2_relu[0][0]']    
                                                                                                  
 conv3_block4_3_bn (BatchNormal  (None, 28, 28, 512)  2048       ['conv3_block4_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv3_block4_add (Add)         (None, 28, 28, 512)  0           ['conv3_block3_out[0][0]',       
                                                                  'conv3_block4_3_bn[0][0]']      
                                                                                                  
 conv3_block4_out (Activation)  (None, 28, 28, 512)  0           ['conv3_block4_add[0][0]']       
                                                                                                  
 conv4_block1_1_conv (Conv2D)   (None, 14, 14, 256)  131328      ['conv3_block4_out[0][0]']       
                                                                                                  
 conv4_block1_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block1_1_relu[0][0]']    
                                                                                                  
 conv4_block1_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block1_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block1_0_conv (Conv2D)   (None, 14, 14, 1024  525312      ['conv3_block4_out[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block1_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block1_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block1_0_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_0_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block1_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block1_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_0_bn[0][0]',      
                                )                                 'conv4_block1_3_bn[0][0]']      
                                                                                                  
 conv4_block1_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block1_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block2_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block1_out[0][0]']       
                                                                                                  
 conv4_block2_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block2_1_relu[0][0]']    
                                                                                                  
 conv4_block2_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block2_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block2_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block2_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block2_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block2_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block2_add (Add)         (None, 14, 14, 1024  0           ['conv4_block1_out[0][0]',       
                                )                                 'conv4_block2_3_bn[0][0]']      
                                                                                                  
 conv4_block2_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block2_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block3_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block2_out[0][0]']       
                                                                                                  
 conv4_block3_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block3_1_relu[0][0]']    
                                                                                                  
 conv4_block3_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block3_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block3_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block3_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block3_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block3_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block3_add (Add)         (None, 14, 14, 1024  0           ['conv4_block2_out[0][0]',       
                                )                                 'conv4_block3_3_bn[0][0]']      
                                                                                                  
 conv4_block3_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block3_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block4_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block3_out[0][0]']       
                                                                                                  
 conv4_block4_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block4_1_relu[0][0]']    
                                                                                                  
 conv4_block4_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block4_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block4_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block4_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block4_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block4_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block4_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block4_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block4_add (Add)         (None, 14, 14, 1024  0           ['conv4_block3_out[0][0]',       
                                )                                 'conv4_block4_3_bn[0][0]']      
                                                                                                  
 conv4_block4_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block4_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block5_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block4_out[0][0]']       
                                                                                                  
 conv4_block5_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block5_1_relu[0][0]']    
                                                                                                  
 conv4_block5_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block5_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block5_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block5_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block5_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block5_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block5_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block5_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block5_add (Add)         (None, 14, 14, 1024  0           ['conv4_block4_out[0][0]',       
                                )                                 'conv4_block5_3_bn[0][0]']      
                                                                                                  
 conv4_block5_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block5_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv4_block6_1_conv (Conv2D)   (None, 14, 14, 256)  262400      ['conv4_block5_out[0][0]']       
                                                                                                  
 conv4_block6_1_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_1_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_2_conv (Conv2D)   (None, 14, 14, 256)  590080      ['conv4_block6_1_relu[0][0]']    
                                                                                                  
 conv4_block6_2_bn (BatchNormal  (None, 14, 14, 256)  1024       ['conv4_block6_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv4_block6_2_relu (Activatio  (None, 14, 14, 256)  0          ['conv4_block6_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv4_block6_3_conv (Conv2D)   (None, 14, 14, 1024  263168      ['conv4_block6_2_relu[0][0]']    
                                )                                                                 
                                                                                                  
 conv4_block6_3_bn (BatchNormal  (None, 14, 14, 1024  4096       ['conv4_block6_3_conv[0][0]']    
 ization)                       )                                                                 
                                                                                                  
 conv4_block6_add (Add)         (None, 14, 14, 1024  0           ['conv4_block5_out[0][0]',       
                                )                                 'conv4_block6_3_bn[0][0]']      
                                                                                                  
 conv4_block6_out (Activation)  (None, 14, 14, 1024  0           ['conv4_block6_add[0][0]']       
                                )                                                                 
                                                                                                  
 conv5_block1_1_conv (Conv2D)   (None, 7, 7, 512)    524800      ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block1_1_relu[0][0]']    
                                                                                                  
 conv5_block1_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block1_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block1_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block1_0_conv (Conv2D)   (None, 7, 7, 2048)   2099200     ['conv4_block6_out[0][0]']       
                                                                                                  
 conv5_block1_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block1_2_relu[0][0]']    
                                                                                                  
 conv5_block1_0_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_0_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block1_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block1_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_0_bn[0][0]',      
                                                                  'conv5_block1_3_bn[0][0]']      
                                                                                                  
 conv5_block1_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block1_add[0][0]']       
                                                                                                  
 conv5_block2_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block1_out[0][0]']       
                                                                                                  
 conv5_block2_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block2_1_relu[0][0]']    
                                                                                                  
 conv5_block2_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block2_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block2_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block2_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block2_2_relu[0][0]']    
                                                                                                  
 conv5_block2_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block2_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block2_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block1_out[0][0]',       
                                                                  'conv5_block2_3_bn[0][0]']      
                                                                                                  
 conv5_block2_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block2_add[0][0]']       
                                                                                                  
 conv5_block3_1_conv (Conv2D)   (None, 7, 7, 512)    1049088     ['conv5_block2_out[0][0]']       
                                                                                                  
 conv5_block3_1_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_1_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_1_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_1_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_2_conv (Conv2D)   (None, 7, 7, 512)    2359808     ['conv5_block3_1_relu[0][0]']    
                                                                                                  
 conv5_block3_2_bn (BatchNormal  (None, 7, 7, 512)   2048        ['conv5_block3_2_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_2_relu (Activatio  (None, 7, 7, 512)   0           ['conv5_block3_2_bn[0][0]']      
 n)                                                                                               
                                                                                                  
 conv5_block3_3_conv (Conv2D)   (None, 7, 7, 2048)   1050624     ['conv5_block3_2_relu[0][0]']    
                                                                                                  
 conv5_block3_3_bn (BatchNormal  (None, 7, 7, 2048)  8192        ['conv5_block3_3_conv[0][0]']    
 ization)                                                                                         
                                                                                                  
 conv5_block3_add (Add)         (None, 7, 7, 2048)   0           ['conv5_block2_out[0][0]',       
                                                                  'conv5_block3_3_bn[0][0]']      
                                                                                                  
 conv5_block3_out (Activation)  (None, 7, 7, 2048)   0           ['conv5_block3_add[0][0]']       
                                                                                                  
 avg_pool (GlobalAveragePooling  (None, 2048)        0           ['conv5_block3_out[0][0]']       
 2D)                                                                                              
                                                                                                  
 predictions (Dense)            (None, 12)           24588       ['avg_pool[0][0]']               
                                                                                                  
==================================================================================================
Total params: 23,612,300
Trainable params: 23,559,180
Non-trainable params: 53,120
__________________________________________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 27ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.5110    0.7704    0.6145       331
anarrhichomenum     0.3707    0.7053    0.4860       319
   brevantherum     0.8571    0.0192    0.0376       312
      dulcamara     0.2825    0.3359    0.3069       259
   herposolanum     0.5217    0.1319    0.2105       273
     holophylla     0.5292    0.4589    0.4915       316
     lasiocarpa     0.8097    0.7994    0.8045       314
      melongena     0.6201    0.3675    0.4615       302
    micracantha     0.2291    0.3746    0.2843       315
         petota     0.2978    0.5225    0.3794       289
        solanum     0.4437    0.2620    0.3295       271
          torva     0.2909    0.1605    0.2069       299

       accuracy                         0.4178      3600
      macro avg     0.4803    0.4090    0.3844      3600
   weighted avg     0.4846    0.4178    0.3902      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*******************************
* Mean metrics across 4 folds *
*******************************

	0
accuracy	0.355208
acanthophora.precision	0.574573
acanthophora.recall	0.581872
acanthophora.f1-score	0.567652
acanthophora.support	298.250000
anarrhichomenum.precision	0.430213
anarrhichomenum.recall	0.703196
anarrhichomenum.f1-score	0.460831
anarrhichomenum.support	318.750000
brevantherum.precision	0.385957
brevantherum.recall	0.249190
brevantherum.f1-score	0.209797
brevantherum.support	288.000000
dulcamara.precision	0.384649
dulcamara.recall	0.195036
dulcamara.f1-score	0.224633
dulcamara.support	295.500000
herposolanum.precision	0.402083
herposolanum.recall	0.315441
herposolanum.f1-score	0.251336
herposolanum.support	285.000000
holophylla.precision	0.516875
holophylla.recall	0.331999
holophylla.f1-score	0.374831
holophylla.support	304.750000
lasiocarpa.precision	0.781612
lasiocarpa.recall	0.609413
lasiocarpa.f1-score	0.627092
lasiocarpa.support	312.500000
melongena.precision	0.567061
melongena.recall	0.313610
melongena.f1-score	0.361499
melongena.support	298.000000
micracantha.precision	0.250282
micracantha.recall	0.306187
micracantha.f1-score	0.253060
micracantha.support	300.000000
petota.precision	0.271986
petota.recall	0.379494
petota.f1-score	0.299666
petota.support	304.500000
solanum.precision	0.458515
solanum.recall	0.106630
solanum.f1-score	0.154527
solanum.support	294.500000
torva.precision	0.347437
torva.recall	0.127967
torva.f1-score	0.166701
torva.support	300.250000
macro avg.precision	0.447604
macro avg.recall	0.351670
macro avg.f1-score	0.329302
macro avg.support	3600.000000
weighted avg.precision	0.450277
weighted avg.recall	0.355208
weighted avg.f1-score	0.331415
weighted avg.support	3600.000000

CPU times: total: 1h 34min 59s
Wall time: 51min 42s

RestNet50 with ImageNet Transfer Learning

%%time

kf = ShuffleSplit(n_splits=CV_FOLDS,
                  test_size=VAL_SIZE,
                  random_state=RANDOM_SEED)

split = 1

resnet50_tl_cv_val_pred = []

for train_index, val_index in kf.split(balanced_training_data):

    resnet50_tl_model = tf.keras.models.Sequential(name="resnet50_imagenet")

    imagenet_resnet50_model= tf.keras.applications.ResNet50(
        include_top=False,
        pooling="max",
        input_shape=(224,224,3),
        classes=NUM_CLASSES,
        weights='imagenet',)

    for layer in imagenet_resnet50_model.layers[:-14]:
        layer.trainable=False


    # Add a custom FC layer for actual learning
    resnet50_tl_model.add(imagenet_resnet50_model)
    resnet50_tl_model.add(tf.keras.layers.Flatten())
    #resnet50_tl_model.add(tf.keras.layers.Dense(4096, activation='relu'))
    #resnet50_tl_model.add(tf.keras.layers.Dense(4096, activation='relu'))
    resnet50_tl_model.add(tf.keras.layers.Dense(NUM_CLASSES, activation='softmax'))

    # Add an optimizer
    resnet50_tl_model.compile(optimizer=tf.optimizers.Adam(learning_rate=0.003),
                        loss='categorical_crossentropy',
                        metrics=['accuracy'])
    #resnet50_tl_model.compile(optimizer=tf.optimizers.SGD(learning_rate=0.003, momentum=0.9),
    #                        loss='categorical_crossentropy',
    #                        metrics=['accuracy'])
        
    training_split_data = balanced_training_data.iloc[train_index]
    val_split_data = balanced_training_data.iloc[val_index]

        # Add a progress bar and save checkpoints
    resnet50_tl_callbacks = [
        create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"resnet50_tl_{TRAINING_RUN_ID}_{split}.h5")),
        tf.keras.callbacks.ProgbarLogger(
            count_mode = 'steps',
            stateful_metrics = None
        ),
        tf.keras.callbacks.EarlyStopping(
            monitor='val_accuracy',
            min_delta=0,
            patience=40,
            verbose=1,
            mode='auto',
            restore_best_weights=True
        )
    ]

    fit_params = {
        "x": training_split_data,
        "epochs": 105,
        "callbacks": resnet50_tl_callbacks,
        "validation_data": val_split_data,
        "steps_per_epoch": 64,
        "validation_steps": 10,
    }

    preproc_func = tf.keras.applications.resnet50.preprocess_input

    with tf.device(TRAINING_DEVICE_NAME):

        resnet50_tl_training_history = train_model(resnet50_tl_model,
                                                   fit_params=fit_params,
                                                   preproc_func=preproc_func)
        
        _, pred_report = evaluate_model(resnet50_tl_model,
                                        resnet50_tl_training_history,
                                        fit_params=fit_params,
                                        preproc_func=preproc_func)
        
        resnet50_tl_cv_val_pred.append(pred_report)
    
    split += 1

calculate_cv_mean_metrics(resnet50_tl_cv_val_pred)

****************************
* Started at 4408.09375... *
****************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 9.7812 - accuracy: 0.1016
Epoch 1: val_accuracy improved from -inf to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 11s 136ms/step - loss: 9.7812 - accuracy: 0.1016 - val_loss: 10.5562 - val_accuracy: 0.1063
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 4.6693 - accuracy: 0.1172
Epoch 2: val_accuracy improved from 0.10625 to 0.11250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 123ms/step - loss: 4.6693 - accuracy: 0.1172 - val_loss: 2.6248 - val_accuracy: 0.1125
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 3.9980 - accuracy: 0.1260
Epoch 3: val_accuracy improved from 0.11250 to 0.13750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 121ms/step - loss: 3.9980 - accuracy: 0.1260 - val_loss: 2.7891 - val_accuracy: 0.1375
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 3.5856 - accuracy: 0.1348
Epoch 4: val_accuracy did not improve from 0.13750
64/64 [==============================] - 7s 112ms/step - loss: 3.5856 - accuracy: 0.1348 - val_loss: 2.5130 - val_accuracy: 0.1312
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 3.5041 - accuracy: 0.1533
Epoch 5: val_accuracy improved from 0.13750 to 0.16250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 121ms/step - loss: 3.5041 - accuracy: 0.1533 - val_loss: 2.6289 - val_accuracy: 0.1625
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 3.1963 - accuracy: 0.1689
Epoch 6: val_accuracy improved from 0.16250 to 0.20625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 121ms/step - loss: 3.1963 - accuracy: 0.1689 - val_loss: 2.4895 - val_accuracy: 0.2062
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 3.1937 - accuracy: 0.1621
Epoch 7: val_accuracy improved from 0.20625 to 0.23750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 123ms/step - loss: 3.1937 - accuracy: 0.1621 - val_loss: 2.3971 - val_accuracy: 0.2375
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.7437 - accuracy: 0.1904
Epoch 8: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 111ms/step - loss: 2.7437 - accuracy: 0.1904 - val_loss: 2.4495 - val_accuracy: 0.2188
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.7180 - accuracy: 0.1660
Epoch 9: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 112ms/step - loss: 2.7180 - accuracy: 0.1660 - val_loss: 2.4630 - val_accuracy: 0.1750
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.6787 - accuracy: 0.1807
Epoch 10: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 114ms/step - loss: 2.6787 - accuracy: 0.1807 - val_loss: 2.3248 - val_accuracy: 0.2250
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.6185 - accuracy: 0.1807
Epoch 11: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 110ms/step - loss: 2.6185 - accuracy: 0.1807 - val_loss: 2.2160 - val_accuracy: 0.2313
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.4152 - accuracy: 0.1797
Epoch 12: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 112ms/step - loss: 2.4152 - accuracy: 0.1797 - val_loss: 2.6172 - val_accuracy: 0.1688
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.3867 - accuracy: 0.2051
Epoch 13: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 113ms/step - loss: 2.3867 - accuracy: 0.2051 - val_loss: 2.4021 - val_accuracy: 0.2188
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.3620 - accuracy: 0.1924
Epoch 14: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 110ms/step - loss: 2.3620 - accuracy: 0.1924 - val_loss: 2.5437 - val_accuracy: 0.1500
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.3576 - accuracy: 0.2041
Epoch 15: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 111ms/step - loss: 2.3576 - accuracy: 0.2041 - val_loss: 2.2172 - val_accuracy: 0.2313
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.2703 - accuracy: 0.2275
Epoch 16: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 112ms/step - loss: 2.2703 - accuracy: 0.2275 - val_loss: 2.3520 - val_accuracy: 0.1937
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.3020 - accuracy: 0.2305
Epoch 17: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 110ms/step - loss: 2.3020 - accuracy: 0.2305 - val_loss: 2.5890 - val_accuracy: 0.2000
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.2257 - accuracy: 0.2080
Epoch 18: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 111ms/step - loss: 2.2257 - accuracy: 0.2080 - val_loss: 2.2904 - val_accuracy: 0.2313
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.2487 - accuracy: 0.2236
Epoch 19: val_accuracy did not improve from 0.23750
64/64 [==============================] - 7s 112ms/step - loss: 2.2487 - accuracy: 0.2236 - val_loss: 2.4351 - val_accuracy: 0.2313
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.2961 - accuracy: 0.2227
Epoch 20: val_accuracy improved from 0.23750 to 0.25000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 121ms/step - loss: 2.2961 - accuracy: 0.2227 - val_loss: 2.5697 - val_accuracy: 0.2500
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.2324 - accuracy: 0.2432
Epoch 21: val_accuracy improved from 0.25000 to 0.28125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.2324 - accuracy: 0.2432 - val_loss: 2.2166 - val_accuracy: 0.2812
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.2168 - accuracy: 0.2500
Epoch 22: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 114ms/step - loss: 2.2168 - accuracy: 0.2500 - val_loss: 2.4867 - val_accuracy: 0.2062
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.2358 - accuracy: 0.2354
Epoch 23: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.2358 - accuracy: 0.2354 - val_loss: 2.6120 - val_accuracy: 0.1750
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.1811 - accuracy: 0.2617
Epoch 24: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 112ms/step - loss: 2.1811 - accuracy: 0.2617 - val_loss: 2.4839 - val_accuracy: 0.2000
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.1387 - accuracy: 0.2520
Epoch 25: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 114ms/step - loss: 2.1387 - accuracy: 0.2520 - val_loss: 2.8331 - val_accuracy: 0.1875
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.1543 - accuracy: 0.2383
Epoch 26: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.1543 - accuracy: 0.2383 - val_loss: 2.2987 - val_accuracy: 0.2750
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.1374 - accuracy: 0.2637
Epoch 27: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.1374 - accuracy: 0.2637 - val_loss: 2.5094 - val_accuracy: 0.2438
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.1392 - accuracy: 0.2744
Epoch 28: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 114ms/step - loss: 2.1392 - accuracy: 0.2744 - val_loss: 2.4446 - val_accuracy: 0.2562
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.1015 - accuracy: 0.2744
Epoch 29: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.1015 - accuracy: 0.2744 - val_loss: 2.2231 - val_accuracy: 0.2688
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.1173 - accuracy: 0.2471
Epoch 30: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.1173 - accuracy: 0.2471 - val_loss: 2.4541 - val_accuracy: 0.2625
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 2.1738 - accuracy: 0.2451
Epoch 31: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 114ms/step - loss: 2.1738 - accuracy: 0.2451 - val_loss: 2.2989 - val_accuracy: 0.2625
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.0966 - accuracy: 0.2617
Epoch 32: val_accuracy did not improve from 0.28125
64/64 [==============================] - 7s 111ms/step - loss: 2.0966 - accuracy: 0.2617 - val_loss: 2.1202 - val_accuracy: 0.2812
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 2.1353 - accuracy: 0.2607
Epoch 33: val_accuracy improved from 0.28125 to 0.30000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 126ms/step - loss: 2.1353 - accuracy: 0.2607 - val_loss: 2.0906 - val_accuracy: 0.3000
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 2.1408 - accuracy: 0.2461
Epoch 34: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 113ms/step - loss: 2.1408 - accuracy: 0.2461 - val_loss: 2.3615 - val_accuracy: 0.2188
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 2.1295 - accuracy: 0.2578
Epoch 35: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 111ms/step - loss: 2.1295 - accuracy: 0.2578 - val_loss: 2.2119 - val_accuracy: 0.2438
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 2.0974 - accuracy: 0.2549
Epoch 36: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 110ms/step - loss: 2.0974 - accuracy: 0.2549 - val_loss: 2.2490 - val_accuracy: 0.2188
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 2.1215 - accuracy: 0.2861
Epoch 37: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 113ms/step - loss: 2.1215 - accuracy: 0.2861 - val_loss: 2.2203 - val_accuracy: 0.2625
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 2.1530 - accuracy: 0.2539
Epoch 38: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 110ms/step - loss: 2.1530 - accuracy: 0.2539 - val_loss: 2.3198 - val_accuracy: 0.2313
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 2.0472 - accuracy: 0.2764
Epoch 39: val_accuracy did not improve from 0.30000
64/64 [==============================] - 7s 110ms/step - loss: 2.0472 - accuracy: 0.2764 - val_loss: 2.2323 - val_accuracy: 0.3000
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 2.0886 - accuracy: 0.2666
Epoch 40: val_accuracy improved from 0.30000 to 0.31250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 128ms/step - loss: 2.0886 - accuracy: 0.2666 - val_loss: 2.0446 - val_accuracy: 0.3125
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 2.0670 - accuracy: 0.2812
Epoch 41: val_accuracy improved from 0.31250 to 0.31875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 119ms/step - loss: 2.0670 - accuracy: 0.2812 - val_loss: 2.0212 - val_accuracy: 0.3187
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 2.0874 - accuracy: 0.2793
Epoch 42: val_accuracy did not improve from 0.31875
64/64 [==============================] - 7s 114ms/step - loss: 2.0874 - accuracy: 0.2793 - val_loss: 2.1085 - val_accuracy: 0.3063
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 2.1161 - accuracy: 0.2559
Epoch 43: val_accuracy did not improve from 0.31875
64/64 [==============================] - 7s 115ms/step - loss: 2.1161 - accuracy: 0.2559 - val_loss: 2.1229 - val_accuracy: 0.2937
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 2.1339 - accuracy: 0.2520
Epoch 44: val_accuracy improved from 0.31875 to 0.33125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.1339 - accuracy: 0.2520 - val_loss: 1.9852 - val_accuracy: 0.3313
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 2.1035 - accuracy: 0.2588
Epoch 45: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 112ms/step - loss: 2.1035 - accuracy: 0.2588 - val_loss: 1.9931 - val_accuracy: 0.2812
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 2.0172 - accuracy: 0.2900
Epoch 46: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 112ms/step - loss: 2.0172 - accuracy: 0.2900 - val_loss: 2.0509 - val_accuracy: 0.3000
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 2.0648 - accuracy: 0.2842
Epoch 47: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 110ms/step - loss: 2.0648 - accuracy: 0.2842 - val_loss: 2.2216 - val_accuracy: 0.2750
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 2.0550 - accuracy: 0.2871
Epoch 48: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 111ms/step - loss: 2.0550 - accuracy: 0.2871 - val_loss: 2.3782 - val_accuracy: 0.2375
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 2.0413 - accuracy: 0.2910
Epoch 49: val_accuracy improved from 0.33125 to 0.40625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_1.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.0413 - accuracy: 0.2910 - val_loss: 1.8583 - val_accuracy: 0.4062
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 2.0602 - accuracy: 0.2803
Epoch 50: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 2.0602 - accuracy: 0.2803 - val_loss: 2.1885 - val_accuracy: 0.2188
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 2.0507 - accuracy: 0.2969
Epoch 51: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0507 - accuracy: 0.2969 - val_loss: 1.9618 - val_accuracy: 0.3000
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 2.0516 - accuracy: 0.2549
Epoch 52: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 2.0516 - accuracy: 0.2549 - val_loss: 2.3482 - val_accuracy: 0.2500
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 2.0233 - accuracy: 0.3164
Epoch 53: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 2.0233 - accuracy: 0.3164 - val_loss: 1.9483 - val_accuracy: 0.3063
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 2.0763 - accuracy: 0.2803
Epoch 54: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0763 - accuracy: 0.2803 - val_loss: 1.9815 - val_accuracy: 0.3187
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 2.0491 - accuracy: 0.2881
Epoch 55: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0491 - accuracy: 0.2881 - val_loss: 1.9879 - val_accuracy: 0.3000
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 2.0247 - accuracy: 0.2803
Epoch 56: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 2.0247 - accuracy: 0.2803 - val_loss: 2.0569 - val_accuracy: 0.2750
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.9914 - accuracy: 0.3115
Epoch 57: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.9914 - accuracy: 0.3115 - val_loss: 2.1460 - val_accuracy: 0.3187
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 2.0872 - accuracy: 0.2852
Epoch 58: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 2.0872 - accuracy: 0.2852 - val_loss: 2.1906 - val_accuracy: 0.2250
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 2.0007 - accuracy: 0.3066
Epoch 59: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 2.0007 - accuracy: 0.3066 - val_loss: 2.5043 - val_accuracy: 0.2438
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 2.0124 - accuracy: 0.2734
Epoch 60: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0124 - accuracy: 0.2734 - val_loss: 2.1974 - val_accuracy: 0.2625
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 2.0005 - accuracy: 0.2861
Epoch 61: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 114ms/step - loss: 2.0005 - accuracy: 0.2861 - val_loss: 2.1705 - val_accuracy: 0.2812
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 2.0755 - accuracy: 0.2598
Epoch 62: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 2.0755 - accuracy: 0.2598 - val_loss: 1.9843 - val_accuracy: 0.3063
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 2.0320 - accuracy: 0.2930
Epoch 63: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0320 - accuracy: 0.2930 - val_loss: 2.5575 - val_accuracy: 0.2000
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 2.0686 - accuracy: 0.2939
Epoch 64: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0686 - accuracy: 0.2939 - val_loss: 1.9575 - val_accuracy: 0.3187
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.9595 - accuracy: 0.3105
Epoch 65: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 1.9595 - accuracy: 0.3105 - val_loss: 2.0668 - val_accuracy: 0.3063
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 2.0012 - accuracy: 0.3008
Epoch 66: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 2.0012 - accuracy: 0.3008 - val_loss: 2.3435 - val_accuracy: 0.3063
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 2.0257 - accuracy: 0.2988
Epoch 67: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 114ms/step - loss: 2.0257 - accuracy: 0.2988 - val_loss: 2.1740 - val_accuracy: 0.2625
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 2.0065 - accuracy: 0.2969
Epoch 68: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 2.0065 - accuracy: 0.2969 - val_loss: 2.3022 - val_accuracy: 0.2438
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.9959 - accuracy: 0.2998
Epoch 69: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9959 - accuracy: 0.2998 - val_loss: 1.9650 - val_accuracy: 0.3187
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 2.0112 - accuracy: 0.2930
Epoch 70: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 2.0112 - accuracy: 0.2930 - val_loss: 2.4239 - val_accuracy: 0.2375
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.9939 - accuracy: 0.2861
Epoch 71: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9939 - accuracy: 0.2861 - val_loss: 2.1238 - val_accuracy: 0.2500
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.9831 - accuracy: 0.2910
Epoch 72: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9831 - accuracy: 0.2910 - val_loss: 2.3316 - val_accuracy: 0.2313
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.9850 - accuracy: 0.3164
Epoch 73: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 1.9850 - accuracy: 0.3164 - val_loss: 2.0942 - val_accuracy: 0.3000
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 2.0228 - accuracy: 0.2803
Epoch 74: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 2.0228 - accuracy: 0.2803 - val_loss: 2.0668 - val_accuracy: 0.2500
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 2.0015 - accuracy: 0.2969
Epoch 75: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 2.0015 - accuracy: 0.2969 - val_loss: 2.0452 - val_accuracy: 0.2875
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.9747 - accuracy: 0.3125
Epoch 76: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9747 - accuracy: 0.3125 - val_loss: 2.3290 - val_accuracy: 0.2562
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.9871 - accuracy: 0.2969
Epoch 77: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.9871 - accuracy: 0.2969 - val_loss: 2.0156 - val_accuracy: 0.2937
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.9793 - accuracy: 0.2939
Epoch 78: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9793 - accuracy: 0.2939 - val_loss: 2.0635 - val_accuracy: 0.3625
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.9785 - accuracy: 0.2969
Epoch 79: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 1.9785 - accuracy: 0.2969 - val_loss: 2.0949 - val_accuracy: 0.2625
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.9570 - accuracy: 0.3193
Epoch 80: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 1.9570 - accuracy: 0.3193 - val_loss: 2.1119 - val_accuracy: 0.3438
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 2.0233 - accuracy: 0.2852
Epoch 81: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 2.0233 - accuracy: 0.2852 - val_loss: 1.9587 - val_accuracy: 0.3125
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.9822 - accuracy: 0.3125
Epoch 82: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 1.9822 - accuracy: 0.3125 - val_loss: 1.7857 - val_accuracy: 0.3938
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.9347 - accuracy: 0.3154
Epoch 83: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.9347 - accuracy: 0.3154 - val_loss: 2.5211 - val_accuracy: 0.2812
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.9297 - accuracy: 0.3262
Epoch 84: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9297 - accuracy: 0.3262 - val_loss: 2.2044 - val_accuracy: 0.2562
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.9434 - accuracy: 0.2949
Epoch 85: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 1.9434 - accuracy: 0.2949 - val_loss: 2.5246 - val_accuracy: 0.2250
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.9994 - accuracy: 0.3008
Epoch 86: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 110ms/step - loss: 1.9994 - accuracy: 0.3008 - val_loss: 2.4070 - val_accuracy: 0.2375
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.9817 - accuracy: 0.2920
Epoch 87: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 113ms/step - loss: 1.9817 - accuracy: 0.2920 - val_loss: 2.2478 - val_accuracy: 0.3000
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.9603 - accuracy: 0.3164
Epoch 88: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 112ms/step - loss: 1.9603 - accuracy: 0.3164 - val_loss: 1.9432 - val_accuracy: 0.3562
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.9784 - accuracy: 0.2861
Epoch 89: val_accuracy did not improve from 0.40625
64/64 [==============================] - 7s 111ms/step - loss: 1.9784 - accuracy: 0.2861 - val_loss: 2.0514 - val_accuracy: 0.2812
Restoring model weights from the end of the best epoch: 49.
Epoch 89: early stopping


 ********* Training time: 1075.171875 s.


*****************
* Model Summary *
*****************
Model: "resnet50_imagenet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 resnet50 (Functional)       (None, 2048)              23587712  
                                                                 
 flatten_4 (Flatten)         (None, 2048)              0         
                                                                 
 dense_4 (Dense)             (None, 12)                24588     
                                                                 
=================================================================
Total params: 23,612,300
Trainable params: 4,494,348
Non-trainable params: 19,117,952
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 29ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.4665    0.5705    0.5133       305
anarrhichomenum     0.4877    0.4907    0.4892       322
   brevantherum     0.0000    0.0000    0.0000       266
      dulcamara     0.1779    0.2637    0.2124       311
   herposolanum     0.1813    0.3405    0.2366       279
     holophylla     0.3361    0.4291    0.3769       282
     lasiocarpa     0.5681    0.6262    0.5957       313
      melongena     0.4839    0.6688    0.5615       314
    micracantha     0.2667    0.1739    0.2105       299
         petota     0.3117    0.5133    0.3879       300
        solanum     0.3765    0.1053    0.1645       304
          torva     0.0000    0.0000    0.0000       305

       accuracy                         0.3539      3600
      macro avg     0.3047    0.3485    0.3124      3600
   weighted avg     0.3104    0.3539    0.3179      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

************************
* Started at 5497.0... *
************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 9.8503 - accuracy: 0.0977
Epoch 1: val_accuracy improved from -inf to 0.08125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 11s 134ms/step - loss: 9.8503 - accuracy: 0.0977 - val_loss: 22.7142 - val_accuracy: 0.0812
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 4.5344 - accuracy: 0.1270
Epoch 2: val_accuracy improved from 0.08125 to 0.08750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 119ms/step - loss: 4.5344 - accuracy: 0.1270 - val_loss: 3.2108 - val_accuracy: 0.0875
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 4.3073 - accuracy: 0.1211
Epoch 3: val_accuracy did not improve from 0.08750
64/64 [==============================] - 7s 114ms/step - loss: 4.3073 - accuracy: 0.1211 - val_loss: 2.6725 - val_accuracy: 0.0875
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 4.0181 - accuracy: 0.1338
Epoch 4: val_accuracy improved from 0.08750 to 0.16250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 122ms/step - loss: 4.0181 - accuracy: 0.1338 - val_loss: 2.7170 - val_accuracy: 0.1625
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 3.4527 - accuracy: 0.1455
Epoch 5: val_accuracy improved from 0.16250 to 0.17500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 118ms/step - loss: 3.4527 - accuracy: 0.1455 - val_loss: 2.5119 - val_accuracy: 0.1750
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 3.5159 - accuracy: 0.1445
Epoch 6: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 114ms/step - loss: 3.5159 - accuracy: 0.1445 - val_loss: 2.6885 - val_accuracy: 0.1750
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 3.2669 - accuracy: 0.1807
Epoch 7: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 112ms/step - loss: 3.2669 - accuracy: 0.1807 - val_loss: 2.4887 - val_accuracy: 0.1562
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 3.0566 - accuracy: 0.1436
Epoch 8: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 111ms/step - loss: 3.0566 - accuracy: 0.1436 - val_loss: 2.9362 - val_accuracy: 0.1437
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.8269 - accuracy: 0.1650
Epoch 9: val_accuracy did not improve from 0.17500
64/64 [==============================] - 7s 113ms/step - loss: 2.8269 - accuracy: 0.1650 - val_loss: 2.7073 - val_accuracy: 0.1437
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.7722 - accuracy: 0.1689
Epoch 10: val_accuracy improved from 0.17500 to 0.24375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.7722 - accuracy: 0.1689 - val_loss: 2.2311 - val_accuracy: 0.2438
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.7109 - accuracy: 0.1816
Epoch 11: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 110ms/step - loss: 2.7109 - accuracy: 0.1816 - val_loss: 2.1393 - val_accuracy: 0.2250
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.6327 - accuracy: 0.1836
Epoch 12: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 113ms/step - loss: 2.6327 - accuracy: 0.1836 - val_loss: 2.4914 - val_accuracy: 0.2250
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.4545 - accuracy: 0.1826
Epoch 13: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 110ms/step - loss: 2.4545 - accuracy: 0.1826 - val_loss: 2.2283 - val_accuracy: 0.2188
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.4535 - accuracy: 0.1953
Epoch 14: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 110ms/step - loss: 2.4535 - accuracy: 0.1953 - val_loss: 2.5832 - val_accuracy: 0.1500
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.2875 - accuracy: 0.2051
Epoch 15: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 113ms/step - loss: 2.2875 - accuracy: 0.2051 - val_loss: 2.6106 - val_accuracy: 0.1562
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.3792 - accuracy: 0.2129
Epoch 16: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 111ms/step - loss: 2.3792 - accuracy: 0.2129 - val_loss: 2.1985 - val_accuracy: 0.2375
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.3074 - accuracy: 0.2080
Epoch 17: val_accuracy did not improve from 0.24375
64/64 [==============================] - 7s 111ms/step - loss: 2.3074 - accuracy: 0.2080 - val_loss: 2.3871 - val_accuracy: 0.2188
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.3485 - accuracy: 0.2002
Epoch 18: val_accuracy improved from 0.24375 to 0.30625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 124ms/step - loss: 2.3485 - accuracy: 0.2002 - val_loss: 2.2747 - val_accuracy: 0.3063
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.2555 - accuracy: 0.2275
Epoch 19: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 2.2555 - accuracy: 0.2275 - val_loss: 2.3419 - val_accuracy: 0.2500
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.2360 - accuracy: 0.2344
Epoch 20: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.2360 - accuracy: 0.2344 - val_loss: 2.2133 - val_accuracy: 0.2562
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.1993 - accuracy: 0.2422
Epoch 21: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 2.1993 - accuracy: 0.2422 - val_loss: 2.0711 - val_accuracy: 0.2750
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.2176 - accuracy: 0.2490
Epoch 22: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.2176 - accuracy: 0.2490 - val_loss: 2.0663 - val_accuracy: 0.2812
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.1976 - accuracy: 0.2471
Epoch 23: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.1976 - accuracy: 0.2471 - val_loss: 2.1059 - val_accuracy: 0.2500
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.2096 - accuracy: 0.2412
Epoch 24: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 2.2096 - accuracy: 0.2412 - val_loss: 2.0671 - val_accuracy: 0.2688
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.2061 - accuracy: 0.2285
Epoch 25: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.2061 - accuracy: 0.2285 - val_loss: 2.3344 - val_accuracy: 0.2500
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.2319 - accuracy: 0.2334
Epoch 26: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.2319 - accuracy: 0.2334 - val_loss: 2.0505 - val_accuracy: 0.2875
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.1662 - accuracy: 0.2354
Epoch 27: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 2.1662 - accuracy: 0.2354 - val_loss: 2.2526 - val_accuracy: 0.2438
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.2488 - accuracy: 0.2393
Epoch 28: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.2488 - accuracy: 0.2393 - val_loss: 2.5025 - val_accuracy: 0.1688
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.0877 - accuracy: 0.2861
Epoch 29: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.0877 - accuracy: 0.2861 - val_loss: 2.9600 - val_accuracy: 0.1562
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.1481 - accuracy: 0.2480
Epoch 30: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 114ms/step - loss: 2.1481 - accuracy: 0.2480 - val_loss: 2.4713 - val_accuracy: 0.2250
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 2.2033 - accuracy: 0.2393
Epoch 31: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 113ms/step - loss: 2.2033 - accuracy: 0.2393 - val_loss: 2.2750 - val_accuracy: 0.2313
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.1308 - accuracy: 0.2490
Epoch 32: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.1308 - accuracy: 0.2490 - val_loss: 2.0739 - val_accuracy: 0.2875
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 2.2186 - accuracy: 0.2061
Epoch 33: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 114ms/step - loss: 2.2186 - accuracy: 0.2061 - val_loss: 2.5815 - val_accuracy: 0.1750
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 2.1283 - accuracy: 0.2412
Epoch 34: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.1283 - accuracy: 0.2412 - val_loss: 2.1554 - val_accuracy: 0.2937
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 2.1071 - accuracy: 0.2705
Epoch 35: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.1071 - accuracy: 0.2705 - val_loss: 2.2109 - val_accuracy: 0.2688
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 2.1297 - accuracy: 0.2832
Epoch 36: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 114ms/step - loss: 2.1297 - accuracy: 0.2832 - val_loss: 2.1049 - val_accuracy: 0.3063
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 2.0634 - accuracy: 0.2803
Epoch 37: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 111ms/step - loss: 2.0634 - accuracy: 0.2803 - val_loss: 2.2719 - val_accuracy: 0.2625
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 2.1403 - accuracy: 0.2686
Epoch 38: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 112ms/step - loss: 2.1403 - accuracy: 0.2686 - val_loss: 2.1409 - val_accuracy: 0.2812
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 2.0373 - accuracy: 0.2998
Epoch 39: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 114ms/step - loss: 2.0373 - accuracy: 0.2998 - val_loss: 2.0417 - val_accuracy: 0.2688
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 2.1194 - accuracy: 0.2646
Epoch 40: val_accuracy did not improve from 0.30625
64/64 [==============================] - 7s 110ms/step - loss: 2.1194 - accuracy: 0.2646 - val_loss: 2.1978 - val_accuracy: 0.2500
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 2.0425 - accuracy: 0.2715
Epoch 41: val_accuracy improved from 0.30625 to 0.31875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 121ms/step - loss: 2.0425 - accuracy: 0.2715 - val_loss: 1.9127 - val_accuracy: 0.3187
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 2.1201 - accuracy: 0.2617
Epoch 42: val_accuracy improved from 0.31875 to 0.33125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 125ms/step - loss: 2.1201 - accuracy: 0.2617 - val_loss: 1.9074 - val_accuracy: 0.3313
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 2.0667 - accuracy: 0.2793
Epoch 43: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 111ms/step - loss: 2.0667 - accuracy: 0.2793 - val_loss: 2.5365 - val_accuracy: 0.1813
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 2.0616 - accuracy: 0.2910
Epoch 44: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 112ms/step - loss: 2.0616 - accuracy: 0.2910 - val_loss: 2.0678 - val_accuracy: 0.2875
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 2.0792 - accuracy: 0.2715
Epoch 45: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 113ms/step - loss: 2.0792 - accuracy: 0.2715 - val_loss: 2.0142 - val_accuracy: 0.2750
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 2.0404 - accuracy: 0.2695
Epoch 46: val_accuracy did not improve from 0.33125
64/64 [==============================] - 7s 110ms/step - loss: 2.0404 - accuracy: 0.2695 - val_loss: 2.0404 - val_accuracy: 0.3000
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 2.1117 - accuracy: 0.2715
Epoch 47: val_accuracy improved from 0.33125 to 0.34375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.1117 - accuracy: 0.2715 - val_loss: 1.9697 - val_accuracy: 0.3438
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 2.0592 - accuracy: 0.2822
Epoch 48: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 115ms/step - loss: 2.0592 - accuracy: 0.2822 - val_loss: 2.0414 - val_accuracy: 0.2688
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 2.1009 - accuracy: 0.2568
Epoch 49: val_accuracy did not improve from 0.34375
64/64 [==============================] - 7s 112ms/step - loss: 2.1009 - accuracy: 0.2568 - val_loss: 2.3856 - val_accuracy: 0.2188
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 2.1028 - accuracy: 0.2529
Epoch 50: val_accuracy improved from 0.34375 to 0.35000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_2.h5
64/64 [==============================] - 8s 120ms/step - loss: 2.1028 - accuracy: 0.2529 - val_loss: 2.0256 - val_accuracy: 0.3500
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 2.0454 - accuracy: 0.3008
Epoch 51: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0454 - accuracy: 0.3008 - val_loss: 1.9409 - val_accuracy: 0.2562
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 2.0632 - accuracy: 0.2705
Epoch 52: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.0632 - accuracy: 0.2705 - val_loss: 1.9821 - val_accuracy: 0.3125
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 2.0574 - accuracy: 0.2666
Epoch 53: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0574 - accuracy: 0.2666 - val_loss: 1.9018 - val_accuracy: 0.3438
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 2.0989 - accuracy: 0.2646
Epoch 54: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0989 - accuracy: 0.2646 - val_loss: 2.3757 - val_accuracy: 0.2250
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 2.0596 - accuracy: 0.2783
Epoch 55: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0596 - accuracy: 0.2783 - val_loss: 2.2695 - val_accuracy: 0.2625
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 2.0369 - accuracy: 0.2881
Epoch 56: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0369 - accuracy: 0.2881 - val_loss: 2.0933 - val_accuracy: 0.2812
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.9764 - accuracy: 0.2891
Epoch 57: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9764 - accuracy: 0.2891 - val_loss: 2.0430 - val_accuracy: 0.3063
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 2.0222 - accuracy: 0.3018
Epoch 58: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.0222 - accuracy: 0.3018 - val_loss: 1.9825 - val_accuracy: 0.3187
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 2.0559 - accuracy: 0.2900
Epoch 59: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0559 - accuracy: 0.2900 - val_loss: 2.1429 - val_accuracy: 0.2688
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.9912 - accuracy: 0.2939
Epoch 60: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 114ms/step - loss: 1.9912 - accuracy: 0.2939 - val_loss: 2.0032 - val_accuracy: 0.3187
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.9638 - accuracy: 0.2959
Epoch 61: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9638 - accuracy: 0.2959 - val_loss: 2.3918 - val_accuracy: 0.2313
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 2.0423 - accuracy: 0.3057
Epoch 62: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0423 - accuracy: 0.3057 - val_loss: 1.9531 - val_accuracy: 0.3187
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 2.0115 - accuracy: 0.2910
Epoch 63: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0115 - accuracy: 0.2910 - val_loss: 2.0715 - val_accuracy: 0.3313
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 2.0376 - accuracy: 0.2959
Epoch 64: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0376 - accuracy: 0.2959 - val_loss: 1.8788 - val_accuracy: 0.3438
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.9894 - accuracy: 0.2910
Epoch 65: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9894 - accuracy: 0.2910 - val_loss: 2.3077 - val_accuracy: 0.2313
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 2.0639 - accuracy: 0.2715
Epoch 66: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0639 - accuracy: 0.2715 - val_loss: 2.3708 - val_accuracy: 0.2313
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 2.0666 - accuracy: 0.2764
Epoch 67: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0666 - accuracy: 0.2764 - val_loss: 2.1583 - val_accuracy: 0.2937
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 2.0124 - accuracy: 0.3076
Epoch 68: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0124 - accuracy: 0.3076 - val_loss: 2.3681 - val_accuracy: 0.2750
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.9758 - accuracy: 0.3057
Epoch 69: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9758 - accuracy: 0.3057 - val_loss: 2.0078 - val_accuracy: 0.2937
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.9811 - accuracy: 0.3047
Epoch 70: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.9811 - accuracy: 0.3047 - val_loss: 2.1495 - val_accuracy: 0.2438
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 2.0719 - accuracy: 0.2881
Epoch 71: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.0719 - accuracy: 0.2881 - val_loss: 2.2064 - val_accuracy: 0.3063
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 2.0287 - accuracy: 0.2861
Epoch 72: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0287 - accuracy: 0.2861 - val_loss: 1.9634 - val_accuracy: 0.3000
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 2.0277 - accuracy: 0.2959
Epoch 73: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0277 - accuracy: 0.2959 - val_loss: 2.0232 - val_accuracy: 0.3187
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.9644 - accuracy: 0.2783
Epoch 74: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.9644 - accuracy: 0.2783 - val_loss: 2.0943 - val_accuracy: 0.3125
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 2.0595 - accuracy: 0.2773
Epoch 75: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0595 - accuracy: 0.2773 - val_loss: 2.1096 - val_accuracy: 0.3125
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.9773 - accuracy: 0.3096
Epoch 76: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9773 - accuracy: 0.3096 - val_loss: 1.9098 - val_accuracy: 0.2812
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 2.0371 - accuracy: 0.2930
Epoch 77: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0371 - accuracy: 0.2930 - val_loss: 1.9801 - val_accuracy: 0.3125
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.9874 - accuracy: 0.3242
Epoch 78: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9874 - accuracy: 0.3242 - val_loss: 2.0968 - val_accuracy: 0.2812
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 1.9779 - accuracy: 0.2900
Epoch 79: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9779 - accuracy: 0.2900 - val_loss: 1.9657 - val_accuracy: 0.3313
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.9844 - accuracy: 0.3027
Epoch 80: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 1.9844 - accuracy: 0.3027 - val_loss: 1.9642 - val_accuracy: 0.3000
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 2.0159 - accuracy: 0.2930
Epoch 81: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0159 - accuracy: 0.2930 - val_loss: 2.0018 - val_accuracy: 0.3313
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.9852 - accuracy: 0.2969
Epoch 82: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9852 - accuracy: 0.2969 - val_loss: 1.8741 - val_accuracy: 0.3375
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 2.0389 - accuracy: 0.2842
Epoch 83: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0389 - accuracy: 0.2842 - val_loss: 2.0593 - val_accuracy: 0.3063
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.9768 - accuracy: 0.3086
Epoch 84: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9768 - accuracy: 0.3086 - val_loss: 2.2043 - val_accuracy: 0.2688
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.9731 - accuracy: 0.2998
Epoch 85: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.9731 - accuracy: 0.2998 - val_loss: 1.8762 - val_accuracy: 0.3375
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.9439 - accuracy: 0.3154
Epoch 86: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9439 - accuracy: 0.3154 - val_loss: 2.4517 - val_accuracy: 0.2125
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 1.9912 - accuracy: 0.2939
Epoch 87: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9912 - accuracy: 0.2939 - val_loss: 2.0583 - val_accuracy: 0.2750
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.9634 - accuracy: 0.3076
Epoch 88: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 1.9634 - accuracy: 0.3076 - val_loss: 2.1731 - val_accuracy: 0.2688
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.9754 - accuracy: 0.2920
Epoch 89: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 1.9754 - accuracy: 0.2920 - val_loss: 1.9530 - val_accuracy: 0.3187
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 1.9596 - accuracy: 0.3047
Epoch 90: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 1.9596 - accuracy: 0.3047 - val_loss: 2.6071 - val_accuracy: 0.2125
Restoring model weights from the end of the best epoch: 50.
Epoch 90: early stopping


 ********* Training time: 1087.09375 s.


*****************
* Model Summary *
*****************
Model: "resnet50_imagenet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 resnet50 (Functional)       (None, 2048)              23587712  
                                                                 
 flatten_5 (Flatten)         (None, 2048)              0         
                                                                 
 dense_5 (Dense)             (None, 12)                24588     
                                                                 
=================================================================
Total params: 23,612,300
Trainable params: 4,494,348
Non-trainable params: 19,117,952
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 28ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.4069    0.4260    0.4162       277
anarrhichomenum     0.4498    0.6540    0.5330       315
   brevantherum     0.2377    0.4723    0.3162       307
      dulcamara     0.3500    0.0236    0.0442       297
   herposolanum     0.6667    0.0139    0.0272       288
     holophylla     0.4138    0.2937    0.3436       286
     lasiocarpa     0.7521    0.2898    0.4184       314
      melongena     0.4367    0.5567    0.4894       291
    micracantha     0.2349    0.2397    0.2373       292
         petota     0.2148    0.6707    0.3254       328
        solanum     0.1809    0.1184    0.1431       304
          torva     0.0000    0.0000    0.0000       301

       accuracy                         0.3175      3600
      macro avg     0.3620    0.3132    0.2745      3600
   weighted avg     0.3608    0.3175    0.2758      3600



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

***************************
* Started at 6597.8125... *
***************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 11.1766 - accuracy: 0.1152
Epoch 1: val_accuracy improved from -inf to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 11s 133ms/step - loss: 11.1766 - accuracy: 0.1152 - val_loss: 43.4558 - val_accuracy: 0.1063
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 3.6561 - accuracy: 0.1084
Epoch 2: val_accuracy did not improve from 0.10625
64/64 [==============================] - 7s 113ms/step - loss: 3.6561 - accuracy: 0.1084 - val_loss: 9.2981 - val_accuracy: 0.1000
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 3.5347 - accuracy: 0.1338
Epoch 3: val_accuracy improved from 0.10625 to 0.16875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 124ms/step - loss: 3.5347 - accuracy: 0.1338 - val_loss: 5.4612 - val_accuracy: 0.1688
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 3.4930 - accuracy: 0.1357
Epoch 4: val_accuracy improved from 0.16875 to 0.18750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 126ms/step - loss: 3.4930 - accuracy: 0.1357 - val_loss: 3.0134 - val_accuracy: 0.1875
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 3.2009 - accuracy: 0.1250
Epoch 5: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 3.2009 - accuracy: 0.1250 - val_loss: 2.7365 - val_accuracy: 0.1375
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.7497 - accuracy: 0.1416
Epoch 6: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 110ms/step - loss: 2.7497 - accuracy: 0.1416 - val_loss: 3.4174 - val_accuracy: 0.1437
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.9649 - accuracy: 0.1602
Epoch 7: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.9649 - accuracy: 0.1602 - val_loss: 2.9283 - val_accuracy: 0.1312
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.7077 - accuracy: 0.1709
Epoch 8: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 112ms/step - loss: 2.7077 - accuracy: 0.1709 - val_loss: 3.2074 - val_accuracy: 0.1562
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.6437 - accuracy: 0.1729
Epoch 9: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 109ms/step - loss: 2.6437 - accuracy: 0.1729 - val_loss: 3.4065 - val_accuracy: 0.1187
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.6017 - accuracy: 0.1973
Epoch 10: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.6017 - accuracy: 0.1973 - val_loss: 2.7921 - val_accuracy: 0.1562
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.5681 - accuracy: 0.2002
Epoch 11: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.5681 - accuracy: 0.2002 - val_loss: 2.8528 - val_accuracy: 0.1562
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.5178 - accuracy: 0.1895
Epoch 12: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 110ms/step - loss: 2.5178 - accuracy: 0.1895 - val_loss: 2.4302 - val_accuracy: 0.1562
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.4702 - accuracy: 0.2070
Epoch 13: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 112ms/step - loss: 2.4702 - accuracy: 0.2070 - val_loss: 2.7332 - val_accuracy: 0.1187
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.4257 - accuracy: 0.2090
Epoch 14: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 112ms/step - loss: 2.4257 - accuracy: 0.2090 - val_loss: 2.6699 - val_accuracy: 0.1562
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.3387 - accuracy: 0.2275
Epoch 15: val_accuracy improved from 0.18750 to 0.21875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 120ms/step - loss: 2.3387 - accuracy: 0.2275 - val_loss: 2.6659 - val_accuracy: 0.2188
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.3468 - accuracy: 0.1963
Epoch 16: val_accuracy did not improve from 0.21875
64/64 [==============================] - 7s 113ms/step - loss: 2.3468 - accuracy: 0.1963 - val_loss: 2.7101 - val_accuracy: 0.2000
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.3172 - accuracy: 0.2109
Epoch 17: val_accuracy did not improve from 0.21875
64/64 [==============================] - 7s 114ms/step - loss: 2.3172 - accuracy: 0.2109 - val_loss: 2.4767 - val_accuracy: 0.1937
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.3300 - accuracy: 0.2051
Epoch 18: val_accuracy did not improve from 0.21875
64/64 [==============================] - 7s 110ms/step - loss: 2.3300 - accuracy: 0.2051 - val_loss: 2.7541 - val_accuracy: 0.1750
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.2564 - accuracy: 0.2314
Epoch 19: val_accuracy improved from 0.21875 to 0.23125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.2564 - accuracy: 0.2314 - val_loss: 2.2739 - val_accuracy: 0.2313
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.3363 - accuracy: 0.1982
Epoch 20: val_accuracy improved from 0.23125 to 0.25000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.3363 - accuracy: 0.1982 - val_loss: 2.1122 - val_accuracy: 0.2500
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.2874 - accuracy: 0.2236
Epoch 21: val_accuracy did not improve from 0.25000
64/64 [==============================] - 7s 111ms/step - loss: 2.2874 - accuracy: 0.2236 - val_loss: 2.3468 - val_accuracy: 0.1875
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.1846 - accuracy: 0.2402
Epoch 22: val_accuracy improved from 0.25000 to 0.26250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 124ms/step - loss: 2.1846 - accuracy: 0.2402 - val_loss: 2.0886 - val_accuracy: 0.2625
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.1934 - accuracy: 0.2148
Epoch 23: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 113ms/step - loss: 2.1934 - accuracy: 0.2148 - val_loss: 2.2539 - val_accuracy: 0.2375
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.2313 - accuracy: 0.2188
Epoch 24: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 109ms/step - loss: 2.2313 - accuracy: 0.2188 - val_loss: 2.8617 - val_accuracy: 0.1187
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.2268 - accuracy: 0.2236
Epoch 25: val_accuracy improved from 0.26250 to 0.33750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.2268 - accuracy: 0.2236 - val_loss: 1.9329 - val_accuracy: 0.3375
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.1691 - accuracy: 0.2432
Epoch 26: val_accuracy did not improve from 0.33750
64/64 [==============================] - 7s 111ms/step - loss: 2.1691 - accuracy: 0.2432 - val_loss: 1.9413 - val_accuracy: 0.3250
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.1059 - accuracy: 0.2646
Epoch 27: val_accuracy did not improve from 0.33750
64/64 [==============================] - 7s 111ms/step - loss: 2.1059 - accuracy: 0.2646 - val_loss: 2.3781 - val_accuracy: 0.2125
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.1711 - accuracy: 0.2334
Epoch 28: val_accuracy did not improve from 0.33750
64/64 [==============================] - 7s 112ms/step - loss: 2.1711 - accuracy: 0.2334 - val_loss: 2.0072 - val_accuracy: 0.2688
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.1386 - accuracy: 0.2559
Epoch 29: val_accuracy did not improve from 0.33750
64/64 [==============================] - 7s 111ms/step - loss: 2.1386 - accuracy: 0.2559 - val_loss: 2.1186 - val_accuracy: 0.2562
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.1060 - accuracy: 0.2637
Epoch 30: val_accuracy improved from 0.33750 to 0.35000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 120ms/step - loss: 2.1060 - accuracy: 0.2637 - val_loss: 1.8518 - val_accuracy: 0.3500
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 2.1088 - accuracy: 0.2510
Epoch 31: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.1088 - accuracy: 0.2510 - val_loss: 2.9913 - val_accuracy: 0.1937
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.1702 - accuracy: 0.2588
Epoch 32: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.1702 - accuracy: 0.2588 - val_loss: 1.9306 - val_accuracy: 0.3125
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 2.0957 - accuracy: 0.2598
Epoch 33: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.0957 - accuracy: 0.2598 - val_loss: 2.2717 - val_accuracy: 0.2625
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 2.1003 - accuracy: 0.2578
Epoch 34: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.1003 - accuracy: 0.2578 - val_loss: 1.9517 - val_accuracy: 0.3500
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 2.1274 - accuracy: 0.2617
Epoch 35: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.1274 - accuracy: 0.2617 - val_loss: 2.6244 - val_accuracy: 0.2062
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 2.1173 - accuracy: 0.2617
Epoch 36: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 2.1173 - accuracy: 0.2617 - val_loss: 2.1521 - val_accuracy: 0.2562
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 2.1531 - accuracy: 0.2646
Epoch 37: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.1531 - accuracy: 0.2646 - val_loss: 2.1594 - val_accuracy: 0.2625
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 2.1526 - accuracy: 0.2686
Epoch 38: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.1526 - accuracy: 0.2686 - val_loss: 2.3501 - val_accuracy: 0.2625
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 2.0852 - accuracy: 0.2695
Epoch 39: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 2.0852 - accuracy: 0.2695 - val_loss: 2.0005 - val_accuracy: 0.3313
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 2.1205 - accuracy: 0.2559
Epoch 40: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.1205 - accuracy: 0.2559 - val_loss: 2.1799 - val_accuracy: 0.3250
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 2.0945 - accuracy: 0.2656
Epoch 41: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 2.0945 - accuracy: 0.2656 - val_loss: 1.9478 - val_accuracy: 0.3125
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 2.0707 - accuracy: 0.2852
Epoch 42: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 110ms/step - loss: 2.0707 - accuracy: 0.2852 - val_loss: 2.1393 - val_accuracy: 0.2438
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 2.0212 - accuracy: 0.2568
Epoch 43: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 113ms/step - loss: 2.0212 - accuracy: 0.2568 - val_loss: 2.4299 - val_accuracy: 0.2688
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 2.0899 - accuracy: 0.2568
Epoch 44: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 111ms/step - loss: 2.0899 - accuracy: 0.2568 - val_loss: 1.9229 - val_accuracy: 0.3000
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 2.0571 - accuracy: 0.2793
Epoch 45: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 109ms/step - loss: 2.0571 - accuracy: 0.2793 - val_loss: 2.1407 - val_accuracy: 0.3125
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 2.0801 - accuracy: 0.2773
Epoch 46: val_accuracy did not improve from 0.35000
64/64 [==============================] - 7s 112ms/step - loss: 2.0801 - accuracy: 0.2773 - val_loss: 1.8921 - val_accuracy: 0.3187
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 2.0618 - accuracy: 0.2646
Epoch 47: val_accuracy improved from 0.35000 to 0.35625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 119ms/step - loss: 2.0618 - accuracy: 0.2646 - val_loss: 1.8386 - val_accuracy: 0.3562
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 2.0397 - accuracy: 0.2861
Epoch 48: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 110ms/step - loss: 2.0397 - accuracy: 0.2861 - val_loss: 1.9780 - val_accuracy: 0.3063
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 2.0289 - accuracy: 0.2803
Epoch 49: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 113ms/step - loss: 2.0289 - accuracy: 0.2803 - val_loss: 2.1016 - val_accuracy: 0.2500
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 2.0642 - accuracy: 0.2676
Epoch 50: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 110ms/step - loss: 2.0642 - accuracy: 0.2676 - val_loss: 2.3822 - val_accuracy: 0.2125
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 2.0214 - accuracy: 0.2783
Epoch 51: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 110ms/step - loss: 2.0214 - accuracy: 0.2783 - val_loss: 2.0041 - val_accuracy: 0.2812
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 2.1384 - accuracy: 0.2471
Epoch 52: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 113ms/step - loss: 2.1384 - accuracy: 0.2471 - val_loss: 2.0910 - val_accuracy: 0.2438
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 2.0261 - accuracy: 0.2871
Epoch 53: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 112ms/step - loss: 2.0261 - accuracy: 0.2871 - val_loss: 2.1626 - val_accuracy: 0.2625
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 2.0727 - accuracy: 0.2744
Epoch 54: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 110ms/step - loss: 2.0727 - accuracy: 0.2744 - val_loss: 2.1077 - val_accuracy: 0.2812
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 2.0129 - accuracy: 0.3066
Epoch 55: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 115ms/step - loss: 2.0129 - accuracy: 0.3066 - val_loss: 1.9317 - val_accuracy: 0.3000
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 2.0506 - accuracy: 0.2695
Epoch 56: val_accuracy did not improve from 0.35625
64/64 [==============================] - 7s 111ms/step - loss: 2.0506 - accuracy: 0.2695 - val_loss: 1.8144 - val_accuracy: 0.3438
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 1.9901 - accuracy: 0.3174
Epoch 57: val_accuracy improved from 0.35625 to 0.40000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 121ms/step - loss: 1.9901 - accuracy: 0.3174 - val_loss: 1.6739 - val_accuracy: 0.4000
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 2.0707 - accuracy: 0.2725
Epoch 58: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 2.0707 - accuracy: 0.2725 - val_loss: 2.0424 - val_accuracy: 0.3187
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 2.0268 - accuracy: 0.2832
Epoch 59: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 2.0268 - accuracy: 0.2832 - val_loss: 1.8625 - val_accuracy: 0.3438
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 2.0100 - accuracy: 0.2705
Epoch 60: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 107ms/step - loss: 2.0100 - accuracy: 0.2705 - val_loss: 2.8113 - val_accuracy: 0.2562
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 2.0457 - accuracy: 0.2979
Epoch 61: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 109ms/step - loss: 2.0457 - accuracy: 0.2979 - val_loss: 2.1269 - val_accuracy: 0.3000
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 2.0108 - accuracy: 0.3047
Epoch 62: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 2.0108 - accuracy: 0.3047 - val_loss: 2.0911 - val_accuracy: 0.2812
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 1.9669 - accuracy: 0.2988
Epoch 63: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 1.9669 - accuracy: 0.2988 - val_loss: 1.8431 - val_accuracy: 0.3375
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 2.0496 - accuracy: 0.2734
Epoch 64: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 2.0496 - accuracy: 0.2734 - val_loss: 2.2768 - val_accuracy: 0.2313
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 1.9811 - accuracy: 0.2969
Epoch 65: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 1.9811 - accuracy: 0.2969 - val_loss: 1.7741 - val_accuracy: 0.3750
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 2.0101 - accuracy: 0.2871
Epoch 66: val_accuracy did not improve from 0.40000
64/64 [==============================] - 7s 108ms/step - loss: 2.0101 - accuracy: 0.2871 - val_loss: 1.8571 - val_accuracy: 0.3313
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 1.9930 - accuracy: 0.2949
Epoch 67: val_accuracy improved from 0.40000 to 0.43125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_3.h5
64/64 [==============================] - 8s 118ms/step - loss: 1.9930 - accuracy: 0.2949 - val_loss: 1.6616 - val_accuracy: 0.4313
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 2.0027 - accuracy: 0.2705
Epoch 68: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 109ms/step - loss: 2.0027 - accuracy: 0.2705 - val_loss: 1.6820 - val_accuracy: 0.4062
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 2.0051 - accuracy: 0.3018
Epoch 69: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 2.0051 - accuracy: 0.3018 - val_loss: 2.0241 - val_accuracy: 0.2875
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.9380 - accuracy: 0.3154
Epoch 70: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 112ms/step - loss: 1.9380 - accuracy: 0.3154 - val_loss: 1.8149 - val_accuracy: 0.4125
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 2.0036 - accuracy: 0.3008
Epoch 71: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 109ms/step - loss: 2.0036 - accuracy: 0.3008 - val_loss: 1.8600 - val_accuracy: 0.3250
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 2.0047 - accuracy: 0.2969
Epoch 72: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 2.0047 - accuracy: 0.2969 - val_loss: 2.2350 - val_accuracy: 0.2313
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 2.0048 - accuracy: 0.3076
Epoch 73: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 2.0048 - accuracy: 0.3076 - val_loss: 1.9028 - val_accuracy: 0.3125
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 1.9865 - accuracy: 0.2979
Epoch 74: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9865 - accuracy: 0.2979 - val_loss: 1.8901 - val_accuracy: 0.2937
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.9857 - accuracy: 0.2900
Epoch 75: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9857 - accuracy: 0.2900 - val_loss: 2.5079 - val_accuracy: 0.2125
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 2.0299 - accuracy: 0.2998
Epoch 76: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 2.0299 - accuracy: 0.2998 - val_loss: 1.9009 - val_accuracy: 0.3438
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.9742 - accuracy: 0.2920
Epoch 77: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9742 - accuracy: 0.2920 - val_loss: 1.9686 - val_accuracy: 0.3250
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.9757 - accuracy: 0.2871
Epoch 78: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9757 - accuracy: 0.2871 - val_loss: 1.7540 - val_accuracy: 0.4000
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 2.0105 - accuracy: 0.3027
Epoch 79: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 2.0105 - accuracy: 0.3027 - val_loss: 1.7444 - val_accuracy: 0.3938
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.9518 - accuracy: 0.3125
Epoch 80: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9518 - accuracy: 0.3125 - val_loss: 2.1504 - val_accuracy: 0.3250
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 2.0446 - accuracy: 0.2744
Epoch 81: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 2.0446 - accuracy: 0.2744 - val_loss: 2.2355 - val_accuracy: 0.2500
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 1.9738 - accuracy: 0.3057
Epoch 82: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 112ms/step - loss: 1.9738 - accuracy: 0.3057 - val_loss: 2.0507 - val_accuracy: 0.2812
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.9975 - accuracy: 0.3037
Epoch 83: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9975 - accuracy: 0.3037 - val_loss: 2.0518 - val_accuracy: 0.2688
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.9553 - accuracy: 0.2949
Epoch 84: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9553 - accuracy: 0.2949 - val_loss: 2.2014 - val_accuracy: 0.2875
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.9551 - accuracy: 0.3232
Epoch 85: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 1.9551 - accuracy: 0.3232 - val_loss: 2.5504 - val_accuracy: 0.2375
Epoch 86/105
64/64 [==============================] - ETA: 0s - loss: 1.9649 - accuracy: 0.3008
Epoch 86: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9649 - accuracy: 0.3008 - val_loss: 1.8823 - val_accuracy: 0.3500
Epoch 87/105
64/64 [==============================] - ETA: 0s - loss: 2.0089 - accuracy: 0.2988
Epoch 87: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 2.0089 - accuracy: 0.2988 - val_loss: 2.0038 - val_accuracy: 0.3063
Epoch 88/105
64/64 [==============================] - ETA: 0s - loss: 1.9362 - accuracy: 0.3105
Epoch 88: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 1.9362 - accuracy: 0.3105 - val_loss: 1.9549 - val_accuracy: 0.3375
Epoch 89/105
64/64 [==============================] - ETA: 0s - loss: 1.9741 - accuracy: 0.2939
Epoch 89: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9741 - accuracy: 0.2939 - val_loss: 1.7678 - val_accuracy: 0.4062
Epoch 90/105
64/64 [==============================] - ETA: 0s - loss: 2.0210 - accuracy: 0.2959
Epoch 90: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 2.0210 - accuracy: 0.2959 - val_loss: 1.9524 - val_accuracy: 0.3313
Epoch 91/105
64/64 [==============================] - ETA: 0s - loss: 1.9242 - accuracy: 0.3340
Epoch 91: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 1.9242 - accuracy: 0.3340 - val_loss: 1.7315 - val_accuracy: 0.3750
Epoch 92/105
64/64 [==============================] - ETA: 0s - loss: 1.9023 - accuracy: 0.3311
Epoch 92: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9023 - accuracy: 0.3311 - val_loss: 2.0749 - val_accuracy: 0.2875
Epoch 93/105
64/64 [==============================] - ETA: 0s - loss: 1.9381 - accuracy: 0.3242
Epoch 93: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9381 - accuracy: 0.3242 - val_loss: 1.7280 - val_accuracy: 0.3938
Epoch 94/105
64/64 [==============================] - ETA: 0s - loss: 1.9719 - accuracy: 0.3027
Epoch 94: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 112ms/step - loss: 1.9719 - accuracy: 0.3027 - val_loss: 1.8577 - val_accuracy: 0.3375
Epoch 95/105
64/64 [==============================] - ETA: 0s - loss: 1.9168 - accuracy: 0.3174
Epoch 95: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9168 - accuracy: 0.3174 - val_loss: 2.0917 - val_accuracy: 0.3187
Epoch 96/105
64/64 [==============================] - ETA: 0s - loss: 1.9590 - accuracy: 0.3232
Epoch 96: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9590 - accuracy: 0.3232 - val_loss: 2.3261 - val_accuracy: 0.2625
Epoch 97/105
64/64 [==============================] - ETA: 0s - loss: 1.9481 - accuracy: 0.3115
Epoch 97: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 113ms/step - loss: 1.9481 - accuracy: 0.3115 - val_loss: 2.0445 - val_accuracy: 0.3000
Epoch 98/105
64/64 [==============================] - ETA: 0s - loss: 1.9976 - accuracy: 0.3057
Epoch 98: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9976 - accuracy: 0.3057 - val_loss: 2.1486 - val_accuracy: 0.2688
Epoch 99/105
64/64 [==============================] - ETA: 0s - loss: 1.8851 - accuracy: 0.3252
Epoch 99: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.8851 - accuracy: 0.3252 - val_loss: 2.1356 - val_accuracy: 0.3063
Epoch 100/105
64/64 [==============================] - ETA: 0s - loss: 1.9671 - accuracy: 0.3164
Epoch 100: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 114ms/step - loss: 1.9671 - accuracy: 0.3164 - val_loss: 1.8085 - val_accuracy: 0.3750
Epoch 101/105
64/64 [==============================] - ETA: 0s - loss: 1.9481 - accuracy: 0.3105
Epoch 101: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9481 - accuracy: 0.3105 - val_loss: 1.8217 - val_accuracy: 0.3187
Epoch 102/105
64/64 [==============================] - ETA: 0s - loss: 1.9560 - accuracy: 0.3066
Epoch 102: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9560 - accuracy: 0.3066 - val_loss: 2.1692 - val_accuracy: 0.2562
Epoch 103/105
64/64 [==============================] - ETA: 0s - loss: 1.9232 - accuracy: 0.3262
Epoch 103: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 112ms/step - loss: 1.9232 - accuracy: 0.3262 - val_loss: 1.9074 - val_accuracy: 0.3438
Epoch 104/105
64/64 [==============================] - ETA: 0s - loss: 1.9298 - accuracy: 0.3164
Epoch 104: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 110ms/step - loss: 1.9298 - accuracy: 0.3164 - val_loss: 1.6815 - val_accuracy: 0.4000
Epoch 105/105
64/64 [==============================] - ETA: 0s - loss: 1.9441 - accuracy: 0.3018
Epoch 105: val_accuracy did not improve from 0.43125
64/64 [==============================] - 7s 111ms/step - loss: 1.9441 - accuracy: 0.3018 - val_loss: 1.9913 - val_accuracy: 0.2750


 ********* Training time: 1258.078125 s.


*****************
* Model Summary *
*****************
Model: "resnet50_imagenet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 resnet50 (Functional)       (None, 2048)              23587712  
                                                                 
 flatten_6 (Flatten)         (None, 2048)              0         
                                                                 
 dense_6 (Dense)             (None, 12)                24588     
                                                                 
=================================================================
Total params: 23,612,300
Trainable params: 4,494,348
Non-trainable params: 19,117,952
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 28ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.5593    0.3536    0.4333       280
anarrhichomenum     0.3430    0.6301    0.4442       319
   brevantherum     0.2435    0.5243    0.3325       267
      dulcamara     0.2500    0.0032    0.0063       315
   herposolanum     0.1957    0.0300    0.0520       300
     holophylla     0.3025    0.2537    0.2760       335
     lasiocarpa     0.7102    0.4045    0.5155       309
      melongena     0.5920    0.4175    0.4897       285
    micracantha     0.1489    0.0238    0.0411       294
         petota     0.1819    0.7209    0.2905       301
        solanum     0.1333    0.0736    0.0948       299
          torva     0.2483    0.1250    0.1663       296

       accuracy                         0.2950      3600
      macro avg     0.3257    0.2967    0.2618      3600
   weighted avg     0.3250    0.2950    0.2605      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

**************************
* Started at 7869.375... *
**************************
Found 14400 validated image filenames belonging to 12 classes.
Found 3600 validated image filenames belonging to 12 classes.
Epoch 1/105
64/64 [==============================] - ETA: 0s - loss: 10.8682 - accuracy: 0.1182
Epoch 1: val_accuracy improved from -inf to 0.10625, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 11s 131ms/step - loss: 10.8682 - accuracy: 0.1182 - val_loss: 27.3072 - val_accuracy: 0.1063
Epoch 2/105
64/64 [==============================] - ETA: 0s - loss: 3.4741 - accuracy: 0.1201
Epoch 2: val_accuracy improved from 0.10625 to 0.12500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 123ms/step - loss: 3.4741 - accuracy: 0.1201 - val_loss: 7.4388 - val_accuracy: 0.1250
Epoch 3/105
64/64 [==============================] - ETA: 0s - loss: 3.5119 - accuracy: 0.1113
Epoch 3: val_accuracy did not improve from 0.12500
64/64 [==============================] - 7s 113ms/step - loss: 3.5119 - accuracy: 0.1113 - val_loss: 4.0268 - val_accuracy: 0.0812
Epoch 4/105
64/64 [==============================] - ETA: 0s - loss: 3.0805 - accuracy: 0.1562
Epoch 4: val_accuracy improved from 0.12500 to 0.15000, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 120ms/step - loss: 3.0805 - accuracy: 0.1562 - val_loss: 2.7984 - val_accuracy: 0.1500
Epoch 5/105
64/64 [==============================] - ETA: 0s - loss: 3.0443 - accuracy: 0.1562
Epoch 5: val_accuracy improved from 0.15000 to 0.16875, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 126ms/step - loss: 3.0443 - accuracy: 0.1562 - val_loss: 3.3048 - val_accuracy: 0.1688
Epoch 6/105
64/64 [==============================] - ETA: 0s - loss: 2.9945 - accuracy: 0.1533
Epoch 6: val_accuracy improved from 0.16875 to 0.17500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.9945 - accuracy: 0.1533 - val_loss: 2.7783 - val_accuracy: 0.1750
Epoch 7/105
64/64 [==============================] - ETA: 0s - loss: 2.8061 - accuracy: 0.1797
Epoch 7: val_accuracy improved from 0.17500 to 0.18750, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.8061 - accuracy: 0.1797 - val_loss: 3.0053 - val_accuracy: 0.1875
Epoch 8/105
64/64 [==============================] - ETA: 0s - loss: 2.7341 - accuracy: 0.1572
Epoch 8: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.7341 - accuracy: 0.1572 - val_loss: 2.9684 - val_accuracy: 0.1625
Epoch 9/105
64/64 [==============================] - ETA: 0s - loss: 2.7743 - accuracy: 0.1543
Epoch 9: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.7743 - accuracy: 0.1543 - val_loss: 2.5333 - val_accuracy: 0.1500
Epoch 10/105
64/64 [==============================] - ETA: 0s - loss: 2.5824 - accuracy: 0.1738
Epoch 10: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 111ms/step - loss: 2.5824 - accuracy: 0.1738 - val_loss: 2.7350 - val_accuracy: 0.1813
Epoch 11/105
64/64 [==============================] - ETA: 0s - loss: 2.6452 - accuracy: 0.1758
Epoch 11: val_accuracy did not improve from 0.18750
64/64 [==============================] - 7s 113ms/step - loss: 2.6452 - accuracy: 0.1758 - val_loss: 2.9581 - val_accuracy: 0.1437
Epoch 12/105
64/64 [==============================] - ETA: 0s - loss: 2.3938 - accuracy: 0.2207
Epoch 12: val_accuracy improved from 0.18750 to 0.19375, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 122ms/step - loss: 2.3938 - accuracy: 0.2207 - val_loss: 2.5608 - val_accuracy: 0.1937
Epoch 13/105
64/64 [==============================] - ETA: 0s - loss: 2.3898 - accuracy: 0.1807
Epoch 13: val_accuracy improved from 0.19375 to 0.26250, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 120ms/step - loss: 2.3898 - accuracy: 0.1807 - val_loss: 2.2575 - val_accuracy: 0.2625
Epoch 14/105
64/64 [==============================] - ETA: 0s - loss: 2.4357 - accuracy: 0.1943
Epoch 14: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 110ms/step - loss: 2.4357 - accuracy: 0.1943 - val_loss: 2.7509 - val_accuracy: 0.1937
Epoch 15/105
64/64 [==============================] - ETA: 0s - loss: 2.3614 - accuracy: 0.1787
Epoch 15: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 112ms/step - loss: 2.3614 - accuracy: 0.1787 - val_loss: 2.4005 - val_accuracy: 0.2188
Epoch 16/105
64/64 [==============================] - ETA: 0s - loss: 2.3143 - accuracy: 0.2129
Epoch 16: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 110ms/step - loss: 2.3143 - accuracy: 0.2129 - val_loss: 2.9854 - val_accuracy: 0.1375
Epoch 17/105
64/64 [==============================] - ETA: 0s - loss: 2.2797 - accuracy: 0.2041
Epoch 17: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 113ms/step - loss: 2.2797 - accuracy: 0.2041 - val_loss: 2.4325 - val_accuracy: 0.1750
Epoch 18/105
64/64 [==============================] - ETA: 0s - loss: 2.2822 - accuracy: 0.2119
Epoch 18: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 112ms/step - loss: 2.2822 - accuracy: 0.2119 - val_loss: 2.4201 - val_accuracy: 0.2250
Epoch 19/105
64/64 [==============================] - ETA: 0s - loss: 2.3200 - accuracy: 0.2012
Epoch 19: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 111ms/step - loss: 2.3200 - accuracy: 0.2012 - val_loss: 2.7219 - val_accuracy: 0.2000
Epoch 20/105
64/64 [==============================] - ETA: 0s - loss: 2.2928 - accuracy: 0.2227
Epoch 20: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 113ms/step - loss: 2.2928 - accuracy: 0.2227 - val_loss: 2.5263 - val_accuracy: 0.2125
Epoch 21/105
64/64 [==============================] - ETA: 0s - loss: 2.2177 - accuracy: 0.2188
Epoch 21: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 112ms/step - loss: 2.2177 - accuracy: 0.2188 - val_loss: 2.5407 - val_accuracy: 0.1562
Epoch 22/105
64/64 [==============================] - ETA: 0s - loss: 2.2042 - accuracy: 0.2188
Epoch 22: val_accuracy did not improve from 0.26250
64/64 [==============================] - 7s 111ms/step - loss: 2.2042 - accuracy: 0.2188 - val_loss: 2.2118 - val_accuracy: 0.2438
Epoch 23/105
64/64 [==============================] - ETA: 0s - loss: 2.2075 - accuracy: 0.2129
Epoch 23: val_accuracy improved from 0.26250 to 0.32500, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 123ms/step - loss: 2.2075 - accuracy: 0.2129 - val_loss: 2.1733 - val_accuracy: 0.3250
Epoch 24/105
64/64 [==============================] - ETA: 0s - loss: 2.1866 - accuracy: 0.2412
Epoch 24: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.1866 - accuracy: 0.2412 - val_loss: 2.1984 - val_accuracy: 0.2250
Epoch 25/105
64/64 [==============================] - ETA: 0s - loss: 2.1947 - accuracy: 0.2314
Epoch 25: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 110ms/step - loss: 2.1947 - accuracy: 0.2314 - val_loss: 2.4856 - val_accuracy: 0.1813
Epoch 26/105
64/64 [==============================] - ETA: 0s - loss: 2.1538 - accuracy: 0.2490
Epoch 26: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.1538 - accuracy: 0.2490 - val_loss: 2.3231 - val_accuracy: 0.2250
Epoch 27/105
64/64 [==============================] - ETA: 0s - loss: 2.1556 - accuracy: 0.2480
Epoch 27: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 114ms/step - loss: 2.1556 - accuracy: 0.2480 - val_loss: 2.4166 - val_accuracy: 0.1875
Epoch 28/105
64/64 [==============================] - ETA: 0s - loss: 2.1622 - accuracy: 0.2402
Epoch 28: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 111ms/step - loss: 2.1622 - accuracy: 0.2402 - val_loss: 2.8099 - val_accuracy: 0.1250
Epoch 29/105
64/64 [==============================] - ETA: 0s - loss: 2.1635 - accuracy: 0.2363
Epoch 29: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.1635 - accuracy: 0.2363 - val_loss: 2.3098 - val_accuracy: 0.1875
Epoch 30/105
64/64 [==============================] - ETA: 0s - loss: 2.1821 - accuracy: 0.2256
Epoch 30: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 111ms/step - loss: 2.1821 - accuracy: 0.2256 - val_loss: 2.0923 - val_accuracy: 0.2688
Epoch 31/105
64/64 [==============================] - ETA: 0s - loss: 2.1271 - accuracy: 0.2480
Epoch 31: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 111ms/step - loss: 2.1271 - accuracy: 0.2480 - val_loss: 2.0655 - val_accuracy: 0.3250
Epoch 32/105
64/64 [==============================] - ETA: 0s - loss: 2.1230 - accuracy: 0.2549
Epoch 32: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 112ms/step - loss: 2.1230 - accuracy: 0.2549 - val_loss: 2.2836 - val_accuracy: 0.2313
Epoch 33/105
64/64 [==============================] - ETA: 0s - loss: 2.1074 - accuracy: 0.2666
Epoch 33: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 114ms/step - loss: 2.1074 - accuracy: 0.2666 - val_loss: 2.6240 - val_accuracy: 0.1813
Epoch 34/105
64/64 [==============================] - ETA: 0s - loss: 2.1653 - accuracy: 0.2598
Epoch 34: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 110ms/step - loss: 2.1653 - accuracy: 0.2598 - val_loss: 2.2683 - val_accuracy: 0.2625
Epoch 35/105
64/64 [==============================] - ETA: 0s - loss: 2.1643 - accuracy: 0.2529
Epoch 35: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.1643 - accuracy: 0.2529 - val_loss: 2.2263 - val_accuracy: 0.2562
Epoch 36/105
64/64 [==============================] - ETA: 0s - loss: 2.0892 - accuracy: 0.2783
Epoch 36: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.0892 - accuracy: 0.2783 - val_loss: 2.0457 - val_accuracy: 0.3187
Epoch 37/105
64/64 [==============================] - ETA: 0s - loss: 2.1103 - accuracy: 0.2529
Epoch 37: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 112ms/step - loss: 2.1103 - accuracy: 0.2529 - val_loss: 2.1039 - val_accuracy: 0.2500
Epoch 38/105
64/64 [==============================] - ETA: 0s - loss: 2.1531 - accuracy: 0.2500
Epoch 38: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 115ms/step - loss: 2.1531 - accuracy: 0.2500 - val_loss: 2.2437 - val_accuracy: 0.2875
Epoch 39/105
64/64 [==============================] - ETA: 0s - loss: 2.1232 - accuracy: 0.2422
Epoch 39: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 114ms/step - loss: 2.1232 - accuracy: 0.2422 - val_loss: 2.7091 - val_accuracy: 0.1625
Epoch 40/105
64/64 [==============================] - ETA: 0s - loss: 2.0978 - accuracy: 0.2471
Epoch 40: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 111ms/step - loss: 2.0978 - accuracy: 0.2471 - val_loss: 2.0094 - val_accuracy: 0.3187
Epoch 41/105
64/64 [==============================] - ETA: 0s - loss: 2.0968 - accuracy: 0.2705
Epoch 41: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 114ms/step - loss: 2.0968 - accuracy: 0.2705 - val_loss: 2.0660 - val_accuracy: 0.2250
Epoch 42/105
64/64 [==============================] - ETA: 0s - loss: 2.0456 - accuracy: 0.2881
Epoch 42: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 113ms/step - loss: 2.0456 - accuracy: 0.2881 - val_loss: 2.3217 - val_accuracy: 0.2688
Epoch 43/105
64/64 [==============================] - ETA: 0s - loss: 2.1143 - accuracy: 0.2451
Epoch 43: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 112ms/step - loss: 2.1143 - accuracy: 0.2451 - val_loss: 2.3168 - val_accuracy: 0.2438
Epoch 44/105
64/64 [==============================] - ETA: 0s - loss: 2.0516 - accuracy: 0.2764
Epoch 44: val_accuracy did not improve from 0.32500
64/64 [==============================] - 7s 114ms/step - loss: 2.0516 - accuracy: 0.2764 - val_loss: 2.5059 - val_accuracy: 0.2438
Epoch 45/105
64/64 [==============================] - ETA: 0s - loss: 2.0698 - accuracy: 0.2744
Epoch 45: val_accuracy improved from 0.32500 to 0.38125, saving model to P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\resnet50_tl_20230216202451_4.h5
64/64 [==============================] - 8s 121ms/step - loss: 2.0698 - accuracy: 0.2744 - val_loss: 1.8414 - val_accuracy: 0.3812
Epoch 46/105
64/64 [==============================] - ETA: 0s - loss: 2.0698 - accuracy: 0.2803
Epoch 46: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0698 - accuracy: 0.2803 - val_loss: 2.4949 - val_accuracy: 0.1937
Epoch 47/105
64/64 [==============================] - ETA: 0s - loss: 2.0292 - accuracy: 0.2646
Epoch 47: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0292 - accuracy: 0.2646 - val_loss: 2.3341 - val_accuracy: 0.2875
Epoch 48/105
64/64 [==============================] - ETA: 0s - loss: 2.0339 - accuracy: 0.2744
Epoch 48: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0339 - accuracy: 0.2744 - val_loss: 2.2678 - val_accuracy: 0.2750
Epoch 49/105
64/64 [==============================] - ETA: 0s - loss: 2.1082 - accuracy: 0.2471
Epoch 49: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.1082 - accuracy: 0.2471 - val_loss: 2.0114 - val_accuracy: 0.3500
Epoch 50/105
64/64 [==============================] - ETA: 0s - loss: 2.1141 - accuracy: 0.2539
Epoch 50: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.1141 - accuracy: 0.2539 - val_loss: 2.8449 - val_accuracy: 0.2688
Epoch 51/105
64/64 [==============================] - ETA: 0s - loss: 2.0665 - accuracy: 0.2842
Epoch 51: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0665 - accuracy: 0.2842 - val_loss: 2.2992 - val_accuracy: 0.2188
Epoch 52/105
64/64 [==============================] - ETA: 0s - loss: 1.9940 - accuracy: 0.2920
Epoch 52: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9940 - accuracy: 0.2920 - val_loss: 2.0285 - val_accuracy: 0.2625
Epoch 53/105
64/64 [==============================] - ETA: 0s - loss: 2.0422 - accuracy: 0.2744
Epoch 53: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 112ms/step - loss: 2.0422 - accuracy: 0.2744 - val_loss: 2.5333 - val_accuracy: 0.2438
Epoch 54/105
64/64 [==============================] - ETA: 0s - loss: 2.0237 - accuracy: 0.2949
Epoch 54: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0237 - accuracy: 0.2949 - val_loss: 2.1719 - val_accuracy: 0.2750
Epoch 55/105
64/64 [==============================] - ETA: 0s - loss: 2.0280 - accuracy: 0.2988
Epoch 55: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0280 - accuracy: 0.2988 - val_loss: 2.1162 - val_accuracy: 0.3000
Epoch 56/105
64/64 [==============================] - ETA: 0s - loss: 2.0238 - accuracy: 0.2939
Epoch 56: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0238 - accuracy: 0.2939 - val_loss: 1.9592 - val_accuracy: 0.3187
Epoch 57/105
64/64 [==============================] - ETA: 0s - loss: 2.0376 - accuracy: 0.3154
Epoch 57: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 112ms/step - loss: 2.0376 - accuracy: 0.3154 - val_loss: 2.0958 - val_accuracy: 0.2562
Epoch 58/105
64/64 [==============================] - ETA: 0s - loss: 2.0085 - accuracy: 0.2979
Epoch 58: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 2.0085 - accuracy: 0.2979 - val_loss: 2.1602 - val_accuracy: 0.2750
Epoch 59/105
64/64 [==============================] - ETA: 0s - loss: 2.0016 - accuracy: 0.2822
Epoch 59: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0016 - accuracy: 0.2822 - val_loss: 2.1826 - val_accuracy: 0.3063
Epoch 60/105
64/64 [==============================] - ETA: 0s - loss: 1.9887 - accuracy: 0.3047
Epoch 60: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9887 - accuracy: 0.3047 - val_loss: 2.2969 - val_accuracy: 0.2500
Epoch 61/105
64/64 [==============================] - ETA: 0s - loss: 1.9692 - accuracy: 0.3135
Epoch 61: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9692 - accuracy: 0.3135 - val_loss: 2.0583 - val_accuracy: 0.2812
Epoch 62/105
64/64 [==============================] - ETA: 0s - loss: 2.0474 - accuracy: 0.2725
Epoch 62: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0474 - accuracy: 0.2725 - val_loss: 2.1223 - val_accuracy: 0.2438
Epoch 63/105
64/64 [==============================] - ETA: 0s - loss: 2.0283 - accuracy: 0.2764
Epoch 63: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 2.0283 - accuracy: 0.2764 - val_loss: 2.1590 - val_accuracy: 0.2937
Epoch 64/105
64/64 [==============================] - ETA: 0s - loss: 2.0425 - accuracy: 0.2812
Epoch 64: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 2.0425 - accuracy: 0.2812 - val_loss: 2.0168 - val_accuracy: 0.3313
Epoch 65/105
64/64 [==============================] - ETA: 0s - loss: 2.0154 - accuracy: 0.2871
Epoch 65: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 112ms/step - loss: 2.0154 - accuracy: 0.2871 - val_loss: 2.0075 - val_accuracy: 0.3187
Epoch 66/105
64/64 [==============================] - ETA: 0s - loss: 1.9926 - accuracy: 0.3008
Epoch 66: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9926 - accuracy: 0.3008 - val_loss: 2.0538 - val_accuracy: 0.2875
Epoch 67/105
64/64 [==============================] - ETA: 0s - loss: 2.0447 - accuracy: 0.2910
Epoch 67: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 2.0447 - accuracy: 0.2910 - val_loss: 2.1081 - val_accuracy: 0.2625
Epoch 68/105
64/64 [==============================] - ETA: 0s - loss: 2.0260 - accuracy: 0.2949
Epoch 68: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0260 - accuracy: 0.2949 - val_loss: 2.0824 - val_accuracy: 0.2812
Epoch 69/105
64/64 [==============================] - ETA: 0s - loss: 1.9872 - accuracy: 0.3105
Epoch 69: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9872 - accuracy: 0.3105 - val_loss: 2.5820 - val_accuracy: 0.2625
Epoch 70/105
64/64 [==============================] - ETA: 0s - loss: 1.9876 - accuracy: 0.3125
Epoch 70: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9876 - accuracy: 0.3125 - val_loss: 2.1607 - val_accuracy: 0.2875
Epoch 71/105
64/64 [==============================] - ETA: 0s - loss: 1.9423 - accuracy: 0.3086
Epoch 71: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 112ms/step - loss: 1.9423 - accuracy: 0.3086 - val_loss: 2.4892 - val_accuracy: 0.2688
Epoch 72/105
64/64 [==============================] - ETA: 0s - loss: 1.9992 - accuracy: 0.3018
Epoch 72: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9992 - accuracy: 0.3018 - val_loss: 2.0616 - val_accuracy: 0.2812
Epoch 73/105
64/64 [==============================] - ETA: 0s - loss: 1.9844 - accuracy: 0.3008
Epoch 73: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9844 - accuracy: 0.3008 - val_loss: 2.1092 - val_accuracy: 0.3063
Epoch 74/105
64/64 [==============================] - ETA: 0s - loss: 2.0088 - accuracy: 0.2979
Epoch 74: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 2.0088 - accuracy: 0.2979 - val_loss: 2.3123 - val_accuracy: 0.2875
Epoch 75/105
64/64 [==============================] - ETA: 0s - loss: 1.9405 - accuracy: 0.3203
Epoch 75: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9405 - accuracy: 0.3203 - val_loss: 2.0611 - val_accuracy: 0.3000
Epoch 76/105
64/64 [==============================] - ETA: 0s - loss: 1.9830 - accuracy: 0.3027
Epoch 76: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9830 - accuracy: 0.3027 - val_loss: 2.0765 - val_accuracy: 0.2688
Epoch 77/105
64/64 [==============================] - ETA: 0s - loss: 1.9775 - accuracy: 0.3174
Epoch 77: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 1.9775 - accuracy: 0.3174 - val_loss: 2.1275 - val_accuracy: 0.2688
Epoch 78/105
64/64 [==============================] - ETA: 0s - loss: 1.9616 - accuracy: 0.3184
Epoch 78: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9616 - accuracy: 0.3184 - val_loss: 2.2481 - val_accuracy: 0.2375
Epoch 79/105
64/64 [==============================] - ETA: 0s - loss: 2.0061 - accuracy: 0.3105
Epoch 79: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 2.0061 - accuracy: 0.3105 - val_loss: 2.1111 - val_accuracy: 0.3187
Epoch 80/105
64/64 [==============================] - ETA: 0s - loss: 1.9687 - accuracy: 0.2842
Epoch 80: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 113ms/step - loss: 1.9687 - accuracy: 0.2842 - val_loss: 2.3901 - val_accuracy: 0.2125
Epoch 81/105
64/64 [==============================] - ETA: 0s - loss: 1.9581 - accuracy: 0.3086
Epoch 81: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 111ms/step - loss: 1.9581 - accuracy: 0.3086 - val_loss: 2.6543 - val_accuracy: 0.2562
Epoch 82/105
64/64 [==============================] - ETA: 0s - loss: 2.0318 - accuracy: 0.2881
Epoch 82: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 2.0318 - accuracy: 0.2881 - val_loss: 2.0451 - val_accuracy: 0.3125
Epoch 83/105
64/64 [==============================] - ETA: 0s - loss: 1.9367 - accuracy: 0.3340
Epoch 83: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 112ms/step - loss: 1.9367 - accuracy: 0.3340 - val_loss: 2.0536 - val_accuracy: 0.3125
Epoch 84/105
64/64 [==============================] - ETA: 0s - loss: 1.9776 - accuracy: 0.3096
Epoch 84: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9776 - accuracy: 0.3096 - val_loss: 2.1760 - val_accuracy: 0.2750
Epoch 85/105
64/64 [==============================] - ETA: 0s - loss: 1.9528 - accuracy: 0.3232
Epoch 85: val_accuracy did not improve from 0.38125
64/64 [==============================] - 7s 110ms/step - loss: 1.9528 - accuracy: 0.3232 - val_loss: 2.2718 - val_accuracy: 0.2625
Restoring model weights from the end of the best epoch: 45.
Epoch 85: early stopping


 ********* Training time: 1027.734375 s.


*****************
* Model Summary *
*****************
Model: "resnet50_imagenet"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 resnet50 (Functional)       (None, 2048)              23587712  
                                                                 
 flatten_7 (Flatten)         (None, 2048)              0         
                                                                 
 dense_7 (Dense)             (None, 12)                24588     
                                                                 
=================================================================
Total params: 23,612,300
Trainable params: 4,494,348
Non-trainable params: 19,117,952
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 3600 validated image filenames belonging to 12 classes.
225/225 [==============================] - 7s 28ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.5238    0.2991    0.3808       331
anarrhichomenum     0.6615    0.4044    0.5019       319
   brevantherum     0.2451    0.4038    0.3051       312
      dulcamara     0.2345    0.2046    0.2186       259
   herposolanum     0.2295    0.0513    0.0838       273
     holophylla     0.2564    0.5411    0.3479       316
     lasiocarpa     0.5431    0.6815    0.6045       314
      melongena     0.4821    0.5364    0.5078       302
    micracantha     0.2091    0.2921    0.2437       315
         petota     0.3618    0.5571    0.4387       289
        solanum     0.4000    0.0443    0.0797       271
          torva     0.2427    0.0836    0.1244       299

       accuracy                         0.3494      3600
      macro avg     0.3658    0.3416    0.3197      3600
   weighted avg     0.3702    0.3494    0.3268      3600



********************
* Confusion Matrix *
********************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

*******************************
* Mean metrics across 4 folds *
*******************************

	0
accuracy	0.328958
acanthophora.precision	0.489129
acanthophora.recall	0.412287
acanthophora.f1-score	0.435882
acanthophora.support	298.250000
anarrhichomenum.precision	0.485494
anarrhichomenum.recall	0.544784
anarrhichomenum.f1-score	0.492074
anarrhichomenum.support	318.750000
brevantherum.precision	0.181580
brevantherum.recall	0.350126
brevantherum.f1-score	0.238469
brevantherum.support	288.000000
dulcamara.precision	0.253097
dulcamara.recall	0.123761
dulcamara.f1-score	0.120356
dulcamara.support	295.500000
herposolanum.precision	0.318281
herposolanum.recall	0.108918
herposolanum.f1-score	0.099920
herposolanum.support	285.000000
holophylla.precision	0.327192
holophylla.recall	0.379414
holophylla.f1-score	0.336098
holophylla.support	304.750000
lasiocarpa.precision	0.643389
lasiocarpa.recall	0.500517
lasiocarpa.f1-score	0.533530
lasiocarpa.support	312.500000
melongena.precision	0.498678
melongena.recall	0.544865
melongena.f1-score	0.512118
melongena.support	298.000000
micracantha.precision	0.214898
micracantha.recall	0.182378
micracantha.f1-score	0.183145
micracantha.support	300.000000
petota.precision	0.267569
petota.recall	0.615522
petota.f1-score	0.360635
petota.support	304.500000
solanum.precision	0.272677
solanum.recall	0.085386
solanum.f1-score	0.120557
solanum.support	294.500000
torva.precision	0.122760
torva.recall	0.052153
torva.f1-score	0.072668
torva.support	300.250000
macro avg.precision	0.339562
macro avg.recall	0.325009
macro avg.f1-score	0.292121
macro avg.support	3600.000000
weighted avg.precision	0.341608
weighted avg.recall	0.328958
weighted avg.f1-score	0.295240
weighted avg.support	3600.000000

CPU times: total: 1h 15min 3s
Wall time: 45min 29s

Hyperparameter search for the best performing model

Given the previous results, we'll move on to do a grid search to find the best hyperparameters for the model.

bal_training_set, bal_val_set = split_balanced_dataset(balanced_training_data, 0.1, is_verbose=False)
print(f">>>Train DF: {bal_training_set.shape}\n   Val DF: {bal_val_set.shape}")

>>>Train DF: (16200, 6)
   Val DF: (1800, 6)

"""
fit(
    x=None,
    y=None,
    batch_size=None,
    epochs=1,
    verbose='auto',
    callbacks=None,
    validation_split=0.0,
    validation_data=None,
    shuffle=True,
    class_weight=None,
    sample_weight=None,
    initial_epoch=0,
    steps_per_epoch=None,
    validation_steps=None,
    validation_batch_size=None,
    validation_freq=1,
    max_queue_size=10,
    workers=1,
    use_multiprocessing=False
)

https://keras.io/guides/keras_tuner/getting_started/
https://keras.io/api/keras_tuner/hyperparameters/#choice-method
https://keras.io/guides/keras_tuner/custom_tuner/

"""
import keras_tuner


class MyBestArchitecture(keras_tuner.HyperModel):

    def build(self, hp):

        # Hyperparameters to tune
        model_pooling = hp.Choice("pooling", 
                                  ["avg", "max"],
                                  default="max")
        learning_rate = hp.Choice("learning_rate",
                                  [0.001, 0.01, 0.003, 0.03, 0.006, 0.06],
                                  default=0.01)
        trainable_layers_limit = hp.Choice("trainable_layers",
                                           [-1, -3, -4, -5],
                                           default=-1)
        optimizer_name = hp.Choice("optimizer",
                                   ["sgd", "adam"],
                                   default="sgd")

        vgg16_tl_model = tf.keras.applications.vgg16.VGG16(
            include_top=True,
            weights='imagenet',
            pooling=model_pooling,
            classifier_activation='softmax',
        )

        # Change the output prediction layer to support the 12 classes instead of the 1000 classes in ImageNet
        output = vgg16_tl_model.layers[-2].output
        predictions = tf.keras.layers.Dense(NUM_CLASSES,
                                            activation="softmax")(output)
        vgg16_tl_model = tf.keras.Model(inputs = vgg16_tl_model.input,
                                        outputs = predictions)

        # Freeze all layers but the FC and output
        for layer in (vgg16_tl_model.layers)[:trainable_layers_limit]:
            layer.trainable = False

        # Add an optimizer
        with hp.conditional_scope("optimizer", ["adam"]):
            if optimizer_name == "adam":
                optimizer = tf.optimizers.Adam(learning_rate=learning_rate)
        with hp.conditional_scope("optimizer", ["sgd"]):
            if optimizer_name == "sgd":
                optimizer = tf.optimizers.SGD(learning_rate=learning_rate)
                
        vgg16_tl_model.compile(optimizer=optimizer,
                               loss='categorical_crossentropy',
                               metrics=['accuracy'])
        
        return vgg16_tl_model
    
    def fit(self, hp, model, x, y, **kwargs):

        # More hyperparameter to tune
        batch_size = hp.Choice("batch_size", [16, 32, 64, 128], default=32)
        
        # Add a progress bar and save checkpoints
        vgg16_tl_callbacks = [
            # create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"vgg16_hyperparam_{TRAINING_RUN_ID}.h5")),
            tf.keras.callbacks.ProgbarLogger(
                count_mode = 'steps',
                stateful_metrics = None
            ),
            tf.keras.callbacks.EarlyStopping(
                monitor='val_accuracy',
                min_delta=0,
                patience=64,
                verbose=1,
                mode='auto',
                restore_best_weights=True
            )
        ]

        preproc_func = tf.keras.applications.vgg16.preprocess_input

        train_data_gen = get_train_generator(preproc_func, bal_training_set, batch_size)
        val_data_gen = get_val_generator(bal_val_set, preproc_func, batch_size)

        kwargs["callbacks"].extend(vgg16_tl_callbacks)

        fit_params = {
                        "x": train_data_gen, 
                        "validation_data": val_data_gen,
                        "steps_per_epoch": 64,
                        "validation_steps": 10,
                        "workers": 10,
                     }
        
        tf.keras.backend.clear_session()

        with tf.device(TRAINING_DEVICE_NAME):
            history = model.fit(**fit_params, **kwargs)

        return history

%%time
random_tuner = keras_tuner.RandomSearch(
    MyBestArchitecture(),
    objective="val_accuracy",
    max_trials=4,
    seed=RANDOM_SEED,
    overwrite=True,
    directory=os.path.join(DATA_ROOT_LOCATION, "vgg16tl_hypertuner"),
    project_name="vgg16tl",
)

random_tuner.search(None, None, epochs=104)

display(random_tuner.search_space_summary())
display(random_tuner.results_summary())

Trial 4 Complete [00h 17m 14s]
val_accuracy: 0.5249999761581421

Best val_accuracy So Far: 0.840624988079071
Total elapsed time: 01h 00m 32s
INFO:tensorflow:Oracle triggered exit
Search space summary
Default search space size: 5
pooling (Choice)
{'default': 'max', 'conditions': [], 'values': ['avg', 'max'], 'ordered': False}
learning_rate (Choice)
{'default': 0.01, 'conditions': [], 'values': [0.001, 0.01, 0.003, 0.03, 0.006, 0.06], 'ordered': True}
trainable_layers (Choice)
{'default': -1, 'conditions': [], 'values': [-1, -3, -4, -5], 'ordered': True}
optimizer (Choice)
{'default': 'sgd', 'conditions': [], 'values': ['sgd', 'adam'], 'ordered': False}
batch_size (Choice)
{'default': 32, 'conditions': [], 'values': [16, 32, 64, 128], 'ordered': True}

None

Results summary
Results in P:\CODE\ITESM\tesis-dataset-downloader\solanum_output\vgg16tl_hypertuner\vgg16tl
Showing 10 best trials
<keras_tuner.engine.objective.Objective object at 0x000001E7A987D9F0>
Trial summary
Hyperparameters:
pooling: max
learning_rate: 0.03
trainable_layers: -1
optimizer: adam
batch_size: 32
Score: 0.840624988079071
Trial summary
Hyperparameters:
pooling: max
learning_rate: 0.003
trainable_layers: -3
optimizer: adam
batch_size: 64
Score: 0.6343749761581421
Trial summary
Hyperparameters:
pooling: max
learning_rate: 0.001
trainable_layers: -4
optimizer: sgd
batch_size: 64
Score: 0.5249999761581421
Trial summary
Hyperparameters:
pooling: max
learning_rate: 0.03
trainable_layers: -1
optimizer: sgd
batch_size: 64
Score: 0.515625

None

CPU times: total: 4h 44min
Wall time: 1h 35s

# Fetch the best model that was found
# Get the top 1 models.
models = random_tuner.get_best_models(num_models=1)
best_model = models[0]
best_model.summary()

print(str(random_tuner.get_best_hyperparameters()[0].values))


val_gen = get_val_generator(bal_val_set,
                            tf.keras.applications.vgg16.preprocess_input,
                            batch_size=32)

val_model_predictions = best_model.predict(val_gen,
                                           #steps=val_gen.n // (val_gen.batch_size + 1)
                                           #steps=10
                                          )

pred_report = print_dataset_prediction_report(val_model_predictions,
                                              val_gen.classes,
                                              get_class_names(val_gen))

Model: "model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_1 (InputLayer)        [(None, 224, 224, 3)]     0         
                                                                 
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense (Dense)               (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 49,164
Non-trainable params: 134,260,544
_________________________________________________________________
{'pooling': 'max', 'learning_rate': 0.03, 'trainable_layers': -1, 'optimizer': 'adam', 'batch_size': 32}
Found 1800 validated image filenames belonging to 12 classes.
57/57 [==============================] - 3s 53ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.6486    0.4800    0.5517       150
anarrhichomenum     0.4354    0.8533    0.5766       150
   brevantherum     0.3684    0.7933    0.5032       150
      dulcamara     1.0000    0.0133    0.0263       150
   herposolanum     0.4957    0.3867    0.4345       150
     holophylla     0.8571    0.0400    0.0764       150
     lasiocarpa     0.3590    0.9933    0.5274       150
      melongena     0.6768    0.4467    0.5382       150
    micracantha     0.5778    0.1733    0.2667       150
         petota     0.7895    0.1000    0.1775       150
        solanum     0.3268    0.7733    0.4594       150
          torva     0.3077    0.0267    0.0491       150

       accuracy                         0.4233      1800
      macro avg     0.5702    0.4233    0.3489      1800
   weighted avg     0.5702    0.4233    0.3489      1800



********************
* Confusion Matrix *
********************

Using best hyperparameters

Preparation of training and validation data

# Let's focus on oversampling to 1500

MIN_NUM_CLASSES = 1500
sections_to_oversample = list(training_count_per_section[training_count_per_section["count"] < MIN_NUM_CLASSES]["section"].values)

oversampler = RandomOverSampler(sampling_strategy={k: MIN_NUM_CLASSES for k in sections_to_oversample},
                                random_state=RANDOM_SEED)

final_training_set = imbsample(oversampler, training_set)

# Shuffle the data
final_training_set = final_training_set.sample(frac=1,
                                               random_state=RANDOM_SEED).reset_index(drop=True)

final_train, final_val = split_balanced_dataset(final_training_set, is_verbose=False)

print_marquee("Training")
# display(final_train.head(10))
display(count_rows_by_column(final_train, "section"))

print_marquee("Validation")
# display(final_train.head(10))
display(count_rows_by_column(final_val, "section"))

************
* Training *
************

	section	count
9	petota	1404
0	acanthophora	1350
1	anarrhichomenum	1350
2	brevantherum	1350
3	dulcamara	1350
4	herposolanum	1350
5	holophylla	1350
6	lasiocarpa	1350
7	melongena	1350
8	micracantha	1350
10	solanum	1350
11	torva	1350

**************
* Validation *
**************

	section	count
9	petota	156
0	acanthophora	150
1	anarrhichomenum	150
2	brevantherum	150
3	dulcamara	150
4	herposolanum	150
5	holophylla	150
6	lasiocarpa	150
7	melongena	150
8	micracantha	150
10	solanum	150
11	torva	150

Training of model using best hyperparameters

"""
Best Params found:

Trial summary
Hyperparameters:
pooling: max
learning_rate: 0.03
trainable_layers: -1
optimizer: adam
batch_size: 32
Score: 0.840624988079071

"""
LEARNING_RATE = 0.03
BATCH_SIZE = 32

def train_using_best_params(train_df,
                            val_df,
                            test_df = None,
                            is_binary_class: bool=False,
                            print_loss_graphs: bool=True,
                            print_model_info: bool=True):

    preproc_func = tf.keras.applications.vgg16.preprocess_input
    num_classes = len(train_df["section"].unique())
    classifier_activation = None if is_binary_class else 'softmax'
    loss = 'binary_crossentropy' if is_binary_class else 'categorical_crossentropy'

    train_data_gen = get_train_generator(preproc_func,
                                        train_df,
                                        batch_size=BATCH_SIZE)
    

    """
     Get the ImageNet weights and prevent the training of the network up until the FC layers
     
      We will include the top even though we will still perform fine-tuning on 
      the last 3 Dense layers (the Fully-Connected layers) because we want to
      have their weights not initialized to random values but to ImageNet weights.

      The reason for this is because during experimentation it was found that
      having the FC layers initialized to random weights prevented the Vgg16
      network from learning at all.

    """
    vgg16_imagenet = tf.keras.applications.vgg16.VGG16(
        include_top=True,
        weights='imagenet',
        input_shape=(224, 224, 3),
    )

    # Freeze all ImageNet layers, we will do fine tuning in new Dense layers
    #  this will allow us to fine tune with Dense layers pre-initialized to ImageNet weights
    for layer in vgg16_imagenet.layers:
        layer.trainable = False
    
    vgg16_tl_model = tf.keras.Sequential(
        [
            # Remove the 1000 ImageNet prediction layer to accomodate the new prediction layers
            *vgg16_imagenet.layers[:-1],
            tf.keras.layers.Dense(num_classes, activation="softmax"),
        ]
    )

    # Add an optimizer
    vgg16_tl_model.compile(optimizer=tf.optimizers.Adam(learning_rate=LEARNING_RATE),
                           loss=loss,
                           metrics=['accuracy'])

    # Add a progress bar and save checkpoints
    vgg16_tl_callbacks = [
        # create_model_checkpoint(os.path.join(DATA_ROOT_LOCATION, f"vgg16_tl_final.h5")),
        tf.keras.callbacks.ProgbarLogger(
            count_mode = 'steps',
            stateful_metrics = None
        ),
        tf.keras.callbacks.EarlyStopping(
            monitor='loss',
            min_delta=0,
            patience=100,
            verbose=0,
            mode='auto',
            restore_best_weights=True
        )
    ]

    fit_params = {
                    "x": train_data_gen,
                    "epochs": 500,
                    "callbacks": vgg16_tl_callbacks,
                    "steps_per_epoch": 64,
                    "batch_size": BATCH_SIZE,
                    "workers": 10,
                    "verbose": 0,
                    "validation_data": get_val_generator(val_df,
                                                         preprocessing_func=preproc_func,
                                                         batch_size=BATCH_SIZE)
                    }
        
    with tf.device(TRAINING_DEVICE_NAME):

        vgg16_tl_training_history = vgg16_tl_model.fit(**fit_params)

        # Pass the raw DataFrame so that the ImageDataGen is generated properly
        #  We plug the test_set in the validation_data field to piggy back on the
        #  code that evaluates how well the final model is predicting output
        fit_params["validation_data"] = test_df if test_df else val_df
        
        
        _, pred_report = evaluate_model(vgg16_tl_model,
                                        vgg16_tl_training_history,
                                        fit_params=fit_params,
                                        preproc_func=preproc_func,
                                        print_loss_graphs=print_loss_graphs,
                                        print_model_info=print_model_info)
    
    return vgg16_tl_model, vgg16_tl_training_history

%%time
        
train_using_best_params(final_train, final_val)

Found 16254 validated image filenames belonging to 12 classes.
Found 1806 validated image filenames belonging to 12 classes.


*****************
* Model Summary *
*****************
Model: "sequential_18"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense_42 (Dense)            (None, 12)                49164     
                                                                 
=================================================================
Total params: 134,309,708
Trainable params: 49,164
Non-trainable params: 134,260,544
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 12 classes.
57/57 [==============================] - 3s 53ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.7778    0.5600    0.6512       150
anarrhichomenum     0.7453    0.8000    0.7717       150
   brevantherum     0.5241    0.6533    0.5816       150
      dulcamara     0.4884    0.1400    0.2176       150
   herposolanum     0.2528    0.7533    0.3786       150
     holophylla     0.0000    0.0000    0.0000       150
     lasiocarpa     0.9867    0.4933    0.6578       150
      melongena     0.0000    0.0000    0.0000       150
    micracantha     0.4966    0.4800    0.4881       150
         petota     0.3043    0.8077    0.4421       156
        solanum     0.3838    0.5067    0.4368       150
          torva     0.5000    0.0933    0.1573       150

       accuracy                         0.4419      1806
      macro avg     0.4550    0.4406    0.3986      1806
   weighted avg     0.4545    0.4419    0.3987      1806



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

CPU times: total: 1h 29min 53s
Wall time: 19min 16s

One vs many classes

We had very bad results for many of the classes, so in an effort to troubleshoot what are the most problematic classes, we'll perform a One-vs-many binary classification approach to see what classes need to be worked on.

%%time

def transform_to_binary_class(df, section_name: str):
    df = df.copy()
    df.loc[df["section"] != section_name, "section"] = f"not-{section_name}"
    return df


for section_name in np.sort(final_train["section"].unique()):

    print_marquee(f"{section_name} vs. many")

    one_vs_many_train_df = transform_to_binary_class(final_train, section_name)
    one_vs_many_val_df = transform_to_binary_class(final_val, section_name)

    train_using_best_params(one_vs_many_train_df,
                            one_vs_many_val_df,
                            is_binary_class=True,
                            print_loss_graphs=False,
                            print_model_info=False)

*************************
* acanthophora vs. many *
*************************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 51ms/step


*************************
* Classification Report *
*************************
                  precision    recall  f1-score   support

    acanthophora     0.5932    0.9333    0.7254       150
not-acanthophora     0.9936    0.9420    0.9671      1656

        accuracy                         0.9413      1806
       macro avg     0.7934    0.9377    0.8463      1806
    weighted avg     0.9604    0.9413    0.9471      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

****************************
* anarrhichomenum vs. many *
****************************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
                     precision    recall  f1-score   support

    anarrhichomenum     0.7029    0.8200    0.7569       150
not-anarrhichomenum     0.9834    0.9686    0.9760      1656

           accuracy                         0.9563      1806
          macro avg     0.8432    0.8943    0.8664      1806
       weighted avg     0.9601    0.9563    0.9578      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

*************************
* brevantherum vs. many *
*************************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 55ms/step


*************************
* Classification Report *
*************************
                  precision    recall  f1-score   support

    brevantherum     1.0000    0.0067    0.0132       150
not-brevantherum     0.9175    1.0000    0.9569      1656

        accuracy                         0.9175      1806
       macro avg     0.9587    0.5033    0.4851      1806
    weighted avg     0.9243    0.9175    0.8786      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**********************
* dulcamara vs. many *
**********************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
               precision    recall  f1-score   support

    dulcamara     0.0000    0.0000    0.0000       150
not-dulcamara     0.9169    1.0000    0.9567      1656

     accuracy                         0.9169      1806
    macro avg     0.4585    0.5000    0.4783      1806
 weighted avg     0.8408    0.9169    0.8772      1806



********************
* Confusion Matrix *
********************

C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Charlie\AppData\Local\Programs\Python\Python310\lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

***********************************
* Confusion Matrix as percentages *
***********************************

*************************
* herposolanum vs. many *
*************************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
                  precision    recall  f1-score   support

    herposolanum     0.4943    0.5733    0.5309       150
not-herposolanum     0.9608    0.9469    0.9538      1656

        accuracy                         0.9158      1806
       macro avg     0.7275    0.7601    0.7423      1806
    weighted avg     0.9220    0.9158    0.9186      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

***********************
* holophylla vs. many *
***********************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 53ms/step


*************************
* Classification Report *
*************************
                precision    recall  f1-score   support

    holophylla     0.4591    0.4867    0.4725       150
not-holophylla     0.9532    0.9481    0.9507      1656

      accuracy                         0.9097      1806
     macro avg     0.7062    0.7174    0.7116      1806
  weighted avg     0.9122    0.9097    0.9109      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

***********************
* lasiocarpa vs. many *
***********************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 53ms/step


*************************
* Classification Report *
*************************
                precision    recall  f1-score   support

    lasiocarpa     0.8679    0.6133    0.7187       150
not-lasiocarpa     0.9659    0.9915    0.9785      1656

      accuracy                         0.9601      1806
     macro avg     0.9169    0.8024    0.8486      1806
  weighted avg     0.9577    0.9601    0.9570      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**********************
* melongena vs. many *
**********************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 55ms/step


*************************
* Classification Report *
*************************
               precision    recall  f1-score   support

    melongena     0.9070    0.2600    0.4041       150
not-melongena     0.9370    0.9976    0.9664      1656

     accuracy                         0.9363      1806
    macro avg     0.9220    0.6288    0.6853      1806
 weighted avg     0.9345    0.9363    0.9197      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

************************
* micracantha vs. many *
************************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

    micracantha     0.8571    0.1600    0.2697       150
not-micracantha     0.9291    0.9976    0.9621      1656

       accuracy                         0.9280      1806
      macro avg     0.8931    0.5788    0.6159      1806
   weighted avg     0.9232    0.9280    0.9046      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

*******************
* petota vs. many *
*******************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
              precision    recall  f1-score   support

  not-petota     0.9412    0.9794    0.9599      1650
      petota     0.6180    0.3526    0.4490       156

    accuracy                         0.9252      1806
   macro avg     0.7796    0.6660    0.7044      1806
weighted avg     0.9133    0.9252    0.9158      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

********************
* solanum vs. many *
********************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 52ms/step


*************************
* Classification Report *
*************************
              precision    recall  f1-score   support

 not-solanum     0.9220    0.9988    0.9588      1656
     solanum     0.8333    0.0667    0.1235       150

    accuracy                         0.9214      1806
   macro avg     0.8776    0.5327    0.5411      1806
weighted avg     0.9146    0.9214    0.8895      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

******************
* torva vs. many *
******************
Found 16254 validated image filenames belonging to 2 classes.
Found 1806 validated image filenames belonging to 2 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1806 validated image filenames belonging to 2 classes.
57/57 [==============================] - 3s 54ms/step


*************************
* Classification Report *
*************************
              precision    recall  f1-score   support

   not-torva     0.9511    0.9269    0.9388      1656
       torva     0.3698    0.4733    0.4152       150

    accuracy                         0.8893      1806
   macro avg     0.6604    0.7001    0.6770      1806
weighted avg     0.9028    0.8893    0.8953      1806



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

CPU times: total: 16h 9min 14s
Wall time: 4h 1min 42s

%%time

def transform_to_binary_class(df, section_name: str):
    df = df.copy()
    df.loc[df["section"] != section_name, "section"] = f"not-{section_name}"
    return df


for section_name in np.sort(final_train["section"].unique()):

    print_marquee(f"{section_name} vs. many")

    one_vs_many_train_df = transform_to_binary_class(final_train, section_name)
    one_vs_many_val_df = transform_to_binary_class(final_val, section_name)

    train_using_best_params(one_vs_many_train_df,
                            one_vs_many_val_df,
                            is_binary_class=True,
                            print_loss_graphs=False,
                            print_model_info=False)

Output hidden; open in https://colab.research.google.com to view.

Removing problematic sections

As we can see there are a couple of sections that are not being recognized by the neural network at all and might be introducing too much noise, let's now train a model without the problematic sections and see how the model behaves, we can evaluate later why the problematic sections are deemed so and propose how we could attack that problem.

For this part, we'll remove the classes that were not identified at all and those that had less than 40% F1-Score.

1. Dulcamara 0.3719
2. Herposolanum 0.3542
3. holophylla 0.0
4. melongena 0.2414
5. solanum 0.3412
6. torva 0.0

1. Brevantherum 0.0132
2. Dulcamara 0.0
3. Micracantha 0.2697
4. Solanum 0.1235

problematic_sections = [
    "dulcamara",
    "herposolanum",
    "holophylla",
    "melongena",
    "solanum",
    "torva"
    ]

print_marquee("Non-problematic training count")
no_problematic_train_df = final_train[~final_train.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_train_df, "section"))

print_marquee("Non-problematic validation count")
no_problematic_val_df = final_val[~final_val.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_val_df, "section"))

**********************************
* Non-problematic training count *
**********************************

	section	count
8	petota	1404
0	acanthophora	1350
1	anarrhichomenum	1350
2	brevantherum	1350
3	dulcamara	1350
4	herposolanum	1350
5	lasiocarpa	1350
6	melongena	1350
7	micracantha	1350
9	solanum	1350

************************************
* Non-problematic validation count *
************************************

	section	count
8	petota	156
0	acanthophora	150
1	anarrhichomenum	150
2	brevantherum	150
3	dulcamara	150
4	herposolanum	150
5	lasiocarpa	150
6	melongena	150
7	micracantha	150
9	solanum	150

%%time

problematic_sections = [
    "brevantherum",
    "dulcamara",
    # "herposolanum",
    # "holophylla",
    # "melongena",
    "solanum",
    "micracantha",
    # "torva"
    ]

print_marquee("Non-problematic training count")
no_problematic_train_df = final_train[~final_train.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_train_df, "section"))

print_marquee("Non-problematic validation count")
no_problematic_val_df = final_val[~final_val.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_val_df, "section"))

train_using_best_params(no_problematic_train_df,
                        no_problematic_val_df,
                        print_loss_graphs=True,
                        print_model_info=False)

**********************************
* Non-problematic training count *
**********************************

	section	count
6	petota	1404
0	acanthophora	1350
1	anarrhichomenum	1350
2	herposolanum	1350
3	holophylla	1350
4	lasiocarpa	1350
5	melongena	1350
7	torva	1350

************************************
* Non-problematic validation count *
************************************

	section	count
6	petota	156
0	acanthophora	150
1	anarrhichomenum	150
2	herposolanum	150
3	holophylla	150
4	lasiocarpa	150
5	melongena	150
7	torva	150

Found 10854 validated image filenames belonging to 8 classes.
Found 1206 validated image filenames belonging to 8 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 1206 validated image filenames belonging to 8 classes.
38/38 [==============================] - 2s 51ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.5403    0.8933    0.6734       150
anarrhichomenum     1.0000    0.0667    0.1250       150
   herposolanum     0.5000    0.6333    0.5588       150
     holophylla     0.6341    0.6933    0.6624       150
     lasiocarpa     0.8943    0.7333    0.8059       150
      melongena     0.9492    0.3733    0.5359       150
         petota     0.3434    0.8718    0.4928       156
          torva     0.7500    0.0800    0.1446       150

       accuracy                         0.5448      1206
      macro avg     0.7014    0.5431    0.4998      1206
   weighted avg     0.6996    0.5448    0.4998      1206



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

CPU times: total: 54min 19s
Wall time: 11min 44s

%%time

problematic_sections = [
    "dulcamara",
    "herposolanum",
    "holophylla",
    "melongena",
    "solanum",
    "torva"
    ]

print_marquee("Non-problematic training count")
no_problematic_train_df = final_train[~final_train.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_train_df, "section"))

print_marquee("Non-problematic validation count")
no_problematic_val_df = final_val[~final_val.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_val_df, "section"))

vgg16_m, vgg16_h = train_using_best_params(no_problematic_train_df,
                        no_problematic_val_df,
                        print_loss_graphs=True,
                        print_model_info=False)

**********************************
* Non-problematic training count *
**********************************

	section	count
5	petota	1404
0	acanthophora	1350
1	anarrhichomenum	1350
2	brevantherum	1350
3	lasiocarpa	1350
4	micracantha	1350

************************************
* Non-problematic validation count *
************************************

	section	count
5	petota	156
0	acanthophora	150
1	anarrhichomenum	150
2	brevantherum	150
3	lasiocarpa	150
4	micracantha	150

Found 8154 validated image filenames belonging to 6 classes.
Found 906 validated image filenames belonging to 6 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 906 validated image filenames belonging to 6 classes.
29/29 [==============================] - 2s 53ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.7333    0.9533    0.8290       150
anarrhichomenum     0.8630    0.8400    0.8514       150
   brevantherum     0.8506    0.8733    0.8618       150
     lasiocarpa     1.0000    0.6067    0.7552       150
    micracantha     0.8661    0.7333    0.7942       150
         petota     0.6943    0.8590    0.7679       156

       accuracy                         0.8113       906
      macro avg     0.8346    0.8109    0.8099       906
   weighted avg     0.8336    0.8113    0.8096       906



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

CPU times: total: 1h 58min 25s
Wall time: 23min 16s

# vgg16_m.save(os.path.join(DATA_ROOT_LOCATION, f"vgg16_tl_best_no_problematic_sections_80acc.h5"))

no_problematic_test_df = test_set[~test_set.section.isin(problematic_sections)]
fit_params = {
                    "epochs": 500,
                    "steps_per_epoch": 64,
                    "batch_size": BATCH_SIZE,
                    "workers": 10,
                    "verbose": 0,
                    "validation_data": no_problematic_test_df
                    }
        
with tf.device(TRAINING_DEVICE_NAME):
    
    _, pred_report = evaluate_model(vgg16_m,
                                    vgg16_h,
                                    fit_params=fit_params,
                                    preproc_func=tf.keras.applications.vgg16.preprocess_input,
                                    print_loss_graphs=True,
                                    print_model_info=True)

*****************
* Model Summary *
*****************
Model: "sequential_6"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense_6 (Dense)             (None, 6)                 24582     
                                                                 
=================================================================
Total params: 134,285,126
Trainable params: 24,582
Non-trainable params: 134,260,544
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 310 validated image filenames belonging to 6 classes.
10/10 [==============================] - 4s 434ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.3333    0.7000    0.4516        10
anarrhichomenum     0.0870    0.2000    0.1212        10
   brevantherum     0.8939    0.7284    0.8027        81
     lasiocarpa     1.0000    0.5556    0.7143        18
    micracantha     0.2174    0.2778    0.2439        18
         petota     0.8743    0.8439    0.8588       173

       accuracy                         0.7387       310
      macro avg     0.5676    0.5509    0.5321       310
   weighted avg     0.8057    0.7387    0.7631       310



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

%%time

problematic_sections = [
    "dulcamara",
    "herposolanum",
    "holophylla",
    "melongena",
    "solanum",
    "torva"
    ]

print_marquee("Non-problematic training count")
no_problematic_train_df = final_train[~final_train.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_train_df, "section"))

print_marquee("Non-problematic validation count")
no_problematic_val_df = final_val[~final_val.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_val_df, "section"))

train_using_best_params(no_problematic_train_df,
                        no_problematic_val_df,
                        print_loss_graphs=True,
                        print_model_info=False)

**********************************
* Non-problematic training count *
**********************************

	section	count
5	petota	1404
0	acanthophora	1350
1	anarrhichomenum	1350
2	brevantherum	1350
3	lasiocarpa	1350
4	micracantha	1350

************************************
* Non-problematic validation count *
************************************

	section	count
5	petota	156
0	acanthophora	150
1	anarrhichomenum	150
2	brevantherum	150
3	lasiocarpa	150
4	micracantha	150

Found 8154 validated image filenames belonging to 6 classes.
Found 906 validated image filenames belonging to 6 classes.


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 906 validated image filenames belonging to 6 classes.
29/29 [==============================] - 2s 50ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.9071    0.8467    0.8759       150
anarrhichomenum     0.5952    1.0000    0.7463       150
   brevantherum     0.9085    0.8600    0.8836       150
     lasiocarpa     0.9489    0.8667    0.9059       150
    micracantha     0.7391    0.9067    0.8144       150
         petota     0.9412    0.3077    0.4638       156

       accuracy                         0.7947       906
      macro avg     0.8400    0.7979    0.7816       906
   weighted avg     0.8407    0.7947    0.7795       906



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

CPU times: total: 2h 5min 47s
Wall time: 25min 1s

%%time

problematic_sections = [
    "dulcamara",
    "herposolanum",
    "holophylla",
    "melongena",
    "solanum",
    "torva"
    ]

print_marquee("Non-problematic training count")
no_problematic_train_df = final_train[~final_train.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_train_df, "section"))

print_marquee("Non-problematic validation count")
no_problematic_val_df = final_val[~final_val.section.isin(problematic_sections)]
display(count_rows_by_column(no_problematic_val_df, "section"))

train_using_best_params(no_problematic_train_df,
                        no_problematic_val_df,
                        print_loss_graphs=True,
                        print_model_info=False)

Found 8154 validated image filenames belonging to 6 classes.
Found 906 validated image filenames belonging to 6 classes.


*****************
* Model Summary *
*****************
Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      
                                                                 
 block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     
                                                                 
 block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         
                                                                 
 block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     
                                                                 
 block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    
                                                                 
 block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         
                                                                 
 block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    
                                                                 
 block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    
                                                                 
 block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         
                                                                 
 block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   
                                                                 
 block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   
                                                                 
 block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         
                                                                 
 block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   
                                                                 
 block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         
                                                                 
 flatten (Flatten)           (None, 25088)             0         
                                                                 
 fc1 (Dense)                 (None, 4096)              102764544 
                                                                 
 fc2 (Dense)                 (None, 4096)              16781312  
                                                                 
 dense (Dense)               (None, 6)                 24582     
                                                                 
=================================================================
Total params: 134,285,126
Trainable params: 24,582
Non-trainable params: 134,260,544
_________________________________________________________________


***************************************
* Validation Dataset Confusion Matrix *
***************************************
Found 906 validated image filenames belonging to 6 classes.
29/29 [==============================] - 2s 49ms/step


*************************
* Classification Report *
*************************
                 precision    recall  f1-score   support

   acanthophora     0.8289    0.8400    0.8344       150
anarrhichomenum     0.8125    0.9533    0.8773       150
   brevantherum     0.9706    0.4400    0.6055       150
     lasiocarpa     0.9583    0.7667    0.8519       150
    micracantha     0.6313    0.8333    0.7184       150
         petota     0.6302    0.7756    0.6954       156

       accuracy                         0.7682       906
      macro avg     0.8053    0.7682    0.7638       906
   weighted avg     0.8042    0.7682    0.7634       906



********************
* Confusion Matrix *
********************

***********************************
* Confusion Matrix as percentages *
***********************************

**************************************
* Train/Val Accuracy and Loss graphs *
**************************************

CPU times: total: 1h 2min 15s
Wall time: 12min 26s